Internal Revenue Service v. CM Holdings, Inc. (In Re CM Holdings, Inc.)

254 B.R. 578, 86 A.F.T.R.2d (RIA) 6470, 2000 U.S. Dist. LEXIS 15321, 2000 WL 1545605

• Court: District Court, D. Delaware
• Decided: October 16, 2000
• Docket: CIV.A. 97-695 MMS
• Status: Published

Opinion

OPINION

SCHWARTZ, Senior District Judge.

TABLE OF CONTENTS

I. INTRODUCTION.581

II. BACKGROUND.

A. General Background of Life Insurance .

1. Whole Life Insurance.

2. Universal Life Policy.

B. Development of the MBL COLI VIII Policy and Plan.

C. Camelot’s Decision to Purchase the MBL COLI VII Plan_

D. Features of Camelot’s COLI VIII Policies and Plan.

1. Amounts Due Each Policy Year (Columns (2), (3), and (4))

2. Sources of Funds to Pay Amounts Due (Columns (5), (6), (7), and (8))

a. Policy Years One through Three.

b. Policy Years Four through Seven.

3. Loan Interest Rate and Differential Crediting Rates.

4. Net Equity (Column (9)).

5. Death Benefit (Columns (10)).

6. Net Pre-Interest Deduction Cash Flow (Column (11)).

7. Interest Deduction Tax Benefit (Column (12)).

8. Net After Interest Deduction Cash Flow (Columns (13)).

III. DISCUSSION.

A. Deductibility of Interest Under § 163 and the Sham Transaction Doctrine

1. General Principles.

2. Sham in Fact.

a. Policy Loans.

(1) Policy Loans Made on First Day of Policy Year.

(2) Risk Characteristics of Loans.

(a) Market, Interest Rates, and Reinvestment Risk.

(b) Marketability or Liquidity Risk.

(c) Volatility, Timing, or Call Risk.

(d) Sector Risk.

(e) Credit or Default Risk.

(f) Maturity Risk.

(g) Legal or Operational Risk.

(3) Artificially High Interest Rates.

(4) of Loans

(5) MBL’s Failure to Restrict Policy Loans During Rehabilitation

(6) Only 40% of Interest Paid in Cash.

b. Loading Dividends.

(1) Magnitude of Loading Charges.

(2) Dividends Not Paid From Surplus.

(3) Dividends Paid At the Beginning of the Policy Year

(4) Sources of Dividends. t — i

Guaranteed. t*H

for Dividends . rH

c. Partial Withdrawals. t — í

d. Annual Premium. tH

e. All Features Together. 03

3. Sham in Substance/Economic Substance Doctrine. 03

a. Objective Economic Substance of Camelot’s COLI Transaction 03

(1) Pre-Interest Deduction Profit. W 03

(a) 20- and 40-Year Illustrations Received by Camelot Prior to and Contemporaneous with its COLI Purchase. LO 03 CD

(b) 81-Year Illustrations. CO 03 CD

(2) Substance of Non-tax Components of Camelot’s COLI VIII 05 1 —

Inside Cash Value 05 l — 1

(b) Tax Free Death Benefits. 05 to

(3) Substance of MBL/Hartford’s of Transaction. 05 cu

What Camelot Got For its Premiums and Loans 05 —3

of Economic Sham 05 —Cl

(1) Business Purpose. 05 CO

(2) of Pre-tax Profitability. 05 l — 1

c. Cases Cited By Camelot Do Not Support COLI VIII Plan Having Economic Substance. 05 ^ H

d. Conclusion. 05 ^ CO

B. Generic or “Abusive” Tax Shelter. 05 ^

C. The I.R.C. § 264(c)(1) Four out of Seven Safe Harbor. 05 Cfi

D. Penalties. 05 ^ <1

1. Substantial Authority Exception Under I.R.C. § 6662(d)(2)(B)© 05 ^ <3

2. Reasonable Cause and Good Faith Exception Under I.R.C. § 6664(c)(1). 05 tfx CO

3. Common Law Issue of First Impression Exception. 05 or bO

IV. CONCLUSION. .653

I. INTRODUCTION

Congress has long favored the life insurance industry by giving favorable tax treatment to two components of life insurance policies. First, a death benefit paid out on a life insurance policy is received by the beneficiary free of income tax. Second, income taxes are deferred, and in some cases never paid, on the cash value built-up inside a life insurance policy. Combined with this favorable tax treatment of life insurance policies, interest on many types of loans, including life insurance policy loans, is tax deductible provided certain statutory criteria are met.

One can readily appreciate that these tax advantages have invited talented actuaries to design life insurance policies which approach becoming tax driven investment vehicles and/or tax shelters, which were never intended by Congress to receive favorable life insurance tax benefits. Over the years, Congress has limited, but not eliminated, these tax advantages in an attempt to curb the use of life insurance policies as investment vehicles. In addition, courts have applied the sham transaction doctrine to deny tax benefits where the life insurance policies have as their sole purpose the creation of tax deductions. Thus, Congress and the courts have stepped in when life insurance policies have crossed the line separating insurance against an untimely death and tax driven or tax sheltering investments.

The Internal Revenue Service (“IRS”) calls upon the Court to determine whether a particular set of life insurance policies has crossed this line. The policies at issue, known as COLI VIII policies, were underwritten by the Mutual Benefit Life Insurance Company (“MBL”). The COLI VIII policies were designed to be owned on a broad base of employees, to be financed through a highly leveraged transaction, and to provide the policyholder with a positive cash flow in every year of the policy. To achieve these design goals, the designers of the COLI VIII policies incorporated several innovative features in an attempt to comply with the Internal Revenue Code governing life insurance. In doing so, the designers obviously inched toward that invisible line which separates true life insurance from tax driven or tax sheltering investments.

Camelot Music, Inc., a wholly owned subsidiary of respondent, CM Holdings, Inc., (collectively “Camelot”), Stipulations of Fact (“Stip.”) ¶ 1, Docket Item (“D.I.”) 64, purchased COLI VIII policies on the lives of 1,430 of its employees. Camelot took out large policy loans to pay the first three annual premiums and claimed interest deductions attributable to these loans for the taxable periods ending August 31, 1991, August 31, 1992, August 31, 1993, for Camelot and ending February 28, 1994, for CM Holdings, Inc. The IRS disallowed $13,836,901 of Camelot’s interest deductions. 1 See Petitioner’s Proposed Findings of Fact and Conclusions of Law and Proposed Order, D.I. 117 at 2.

Both Camelot and MBL met financial misfortune unrelated to the COLI VIII policies. MBL went into rehabilitation and its COLI business was ultimately acquired by the Hartford Life Insurance Company (“Hartford”). On August 9, 1996, Camelot filed a voluntary petition under Chapter 11 of the Bankruptcy Code. The IRS filed proofs of claim against Camelot based in part upon its proposed disal-lowance of the deductions taken by Camelot for interest it claims to have paid on policy loans under its MBL COLI VIII policy plan for taxable years 1991 through February, 1994. Stip. ¶¶ 5, 7; J134, J135, J140. 2 On October 15, 1997, Camelot filed an objection to the IRS’s proofs of claim, creating an adversary proceeding. By order dated May 29, 1998, the IRS’s motion for withdrawal of reference was granted with respect to the adversary proceeding. See Internal Revenue Serv. v. CM Holdings, Inc., 221 B.R. 715, 724 (D.Del.1998) (“CM Holdings I”); D.I. 11. This Court has jurisdiction pursuant to 28 U.S.C. §§ 157(d) and 1334.

Trial was held between March 27, 2000, and May 5, 2000, generating 4,994 pages of transcript and 622 admitted exhibits. The Court has considered the Stipulations of Fact and carefully assessed the testimony, demeanor, and credibility of all witnesses, including those who testified by deposition. 3 In addition, the Court has read and evaluated the combined 378 pages of proposed findings of fact, conclusions of law and legal arguments advanced by the respective parties, and has scrutinized all admitted exhibits cited by either litigant in their post trial submissions.

The Court holds the interest deductions should be disallowed for two alternative reasons. First, they should be disallowed because they were created as part of a transaction which was partially a factual sham and, therefore, did not fall within the safe harbor for policy loan interest deductions of Internal Revenue Code (“I.R.C.”) § 264, 26 U.S.C. § 264. Second, the interest is not deductible, pursuant to I.R.C. § 163(a), 26 U.S.C. § 163(a), because the deductions were created as part of a transaction which was a sham in substance. In addition, the Court concludes this is an appropriate case for assessment of accuracy related penalties for substantial understatement of income tax, pursuant to 26 U.S.C. § 6662.

This opinion shall constitute the Court’s Findings of Facts and Conclusions of Law pursuant to Rule 52 of the Federal Rules of Civil Procedure.

II. BACKGROUND

The IRS has challenged Camelot’s federal income tax deductions for interest it allegedly paid on policy loans as part of a highly structured, leveraged transaction to purchase the COLI VIII policies (the transaction will be referred to as the “COLI VIII plan”). This section discusses life insurance generally, the history of the development of the MBL COLI VIII policy, Camelot’s purchase of the COLI VIII policies, and the features of Camelot’s COLI VIII policies and plan. Unfortunately, because of the relatively recent development of the COLI VIII plan, the industry and trial witnesses have used “the same terms to mean different things and different terms to mean the same thing.” Christian DesRochers (“DesRochers”), Tr.1979. Given this state of affairs, in working with the record, the Court has proceeded with the utmost caution.

A. General Background of Life Insurance

All life insurance policies have certain features in common. A life insurance policy is a contract between a life insurance company and a person who purchases the policy, known as the “policyholder.” Glossary of Life Insurance Terms, Appendix F to Petitioner’s Reply Brief, D.I. 127; Glossary of Insurance Terms, Appendix to Respondent’s Post Trial Brief, D.I. 123. 4 The contract insures against the death of an individual, known as the “insured.” Also named in the policy is a “beneficiary,” who receives money from the insurance company upon the death of the insured. The policyholder, insured, and beneficiary may be the same or different persons. The money received upon the death of the insured is known as the “death benefit” and is received free of federal income tax. Jack Rogers (“Rogers”), Tr. 499. The amount of the death benefit is the “face amount” of the policy.

In order to purchase the death benefit, the policyholder pays the insurance company a fee called a premium. EDWARD E. GRAVES ED., MCGILL’S LIFE INSURANCE 26 (2d ed. 1998) (“MCGILL’S LIFE INSURANCE”). 5 The premium is used, in whole or in part, to pay the cost of insurance (“COI”) charge, which is determined by multiplying an assumed rate of mortality times the face amount of the policy. Id. at 29-30, and 34 n. 12. The assumed rate of mortality varies by gender and tends to rise dramatically with age. The duration of the policy, known as the “term,” may vary between one year and the whole life of the insured. Id. at 28.

There are five basic types of life insurance contracts: term, whole life, universal life, endowment, and annuity. Id. at 35. Only whole life and universal life are of interest in this case.

1. Whole Life Insurance

A whole life insurance policy has a term equal to the whole life of the insured. MCGILL’S LIFE INSURANCE at 49. The policyholder must be able to cover the COI charges, which rise dramatically with age, for the entire life of the individual. Id. at 50. Rather than paying increasing premiums each year, a policyholder may pay a level premium over the entire life of the insured or for a limited period of time (limited-payment life insurance). Id. at 49, 57-59. Either way, the premiums are much higher than the COI charges in early years and lower than the COI charges in later years. Id. at 50-51, 57-59.

Every time a premium is paid, the amount of the premium, less an expense charge, is added to the “policy value,” also known as the “cash value.” Id. at 51-52. The expense charge is a percentage of the premium which is set aside to cover the administrative costs of the policy, including broker commissions. Ordinarily, the magnitude of the expense charge exceeds the projected actual expenses by a small amount known as a “margin.” A major reason for the margin is to protect the insurance company against higher than anticipated expenses. Dwight K. Bartlett (“Bartlett”), Tr. 3364, 3374-75; Dan R. McGill (“McGill”), Tr. 3621-24.

In addition, the insurance company credits the policy value with interest, at an interest rate which the parties have labeled as the “crediting rate.” Among other things the crediting rate allows the policy value to grow to support higher future COI charges. The litigants have adopted what the Court understands to be life insurance industry terminology and called this growth “inside build-up.” The inside build-up is tax deferred for federal income tax purposes.

One important feature of whole life policies is the ability of the policyholder to take a “policy loan,” in an amount not exceeding the policy value, based upon ownership of the life insurance policy. MCGILL’S LIFE INSURANCE at 52-54. The term “policy loan” is at best a misnomer and at worst misleading. The policyholder does not borrow its own money from the policy value. Id. Rather, the policyholder borrows directly from the insurance company, using the policy value as collateral. Id. The Court understands the government and Camelot are in agreement that the policy value becomes “encumbered” to the extent of the principal balance and accrued interest on the loan. However, the encumbered policy value continues to be credited with interest which increases the inside build-up. Id. In order for policy loans to be fully secured by the policy value, the total amount of all outstanding policy loans, plus accrued but unpaid loan interest, cannot exceed the policy value. Id.

The interest rate on the policy loan, known as the “policy loan interest rate,” is set in the policy and is either a fixed rate or a variable rate tied by formula to some specified index. Id. at 52. There is no fixed repayment schedule for the policy loan, but all policy loan principal and accrued interest will be deducted from the payment of the death benefit. Id. at 53. Therefore, another way to think of a policy loan is as an advance on the death benefit. Id. If there is sufficient policy value, policy loans may be used by the policyholder to pay premiums. Id.

Another important feature of whole life policies is a nonforfeiture or surrender option. Id. at 54-55. A policy may be surrendered or terminated at any time in exchange for the insurance company paying the policyholder the policy value. Id. at 54. However, at that time, any outstanding policy loans plus accrued interest must be repaid from the policy value. Id. The “cash surrender value,” also known as “net equity,” is the amount of actual cash the policyholder will receive upon surrender of the policy. Steven A. Eisenberg (“Eisenberg”), Tr. 2218-19.

A final notable feature of whole life insurance policies is the payment of “dividends.” MCGILL’S LIFE INSURANCE at 57. There are two types of whole life policies, participating and non-participating. Id. With a participating policy, the policyholder may receive a dividend paid by the insurance company based on the insurance company’s experience. Id. Dividends may be paid to the policyholder in cash, added to policy value, or used to purchase paid up additions to the face amount of the policy. Id. A dividend is declared by the board of directors from a portion of the insurance company’s surplus which government witnesses have called “divisible surplus.” That portion of the company’s surplus labeled “divisible surplus” is also determined by the board of directors for any given year.

2. Universal Life Policy

Universal life was “introduced in 1979 as a revolutionary new ... variation of whole life insurance.” MCGILL’S LIFE INSURANCE at 77. A traditional whole life policy is opaque or, in the words of two witnesses, a “black box.” James M. Cam-pisi (“Campisi”), Tr. 1036; Eisenberg, Tr. 2190. In contrast, a universal life policy is translucent — all the features of the policy are unbundled to allow the policyholder to see how the money is used. Eisenberg, Tr. 2189-90. A universal life insurance policy contains all of the features of a traditional whole life insurance policy plus several additions.

First, universal life offers the ability to pay flexible, as well as fixed, annual premiums. MCGILL’S LIFE INSURANCE at 79; McGill, G737 at 7. Of course, over the life of the insured, the total policy value must still be able to support payment of the COI charges. MCGILL’S LIFE INSURANCE at 50, 80. However, this feature allows the policyholder to maintain either a very high or very low policy value, depending on his investment goals. Id. at 79-80.

Second, universal life offers a choice between fixed or increasing death benefits. Id. at 83-84; McGill, G737 at 8. In an increasing death benefit policy, the death benefit may increase as the policy value increases. MCGILL’S LIFE INSURANCE at 83-84.

Third, universal life permits the policyholder to make a partial withdrawal of a portion of the policy value, without having the amount withdrawn treated as a policy loan. McGill, G737 at 8. A partial withdrawal has two potentially negative consequences: it reduces the death benefit and it may negatively affect future earnings as there is less policy value to accumulate inside build-up. MCGILL’S LIFE INSURANCE at 84.

Finally, and most importantly, universal life pioneered the use of a “differential crediting rate.” MCGILL’S LIFE INSURANCE at 84-87. As previously rehearsed, a traditional whole life policy has two interest rates: the policy loan interest rate and the crediting rate on the policy value. In contrast, a universal life policy has three interest rates: (1) a policy loan interest rate; (2) a “loaned crediting rate,” an interest rate credited to the portion of the policy value equal to the policy loan balance (i.e., the “encumbered policy value”); and, (3) an “unloaned crediting rate,” an interest rate credited to the remaining portion of policy value (i.e., the “unencumbered policy value”).

The loaned crediting rate is usually less than the unloaned crediting rate in an effort to avoid the effects of disintermediation. MCGILL’S LIFE INSURANCE at 84-87. Disintermediation occurred, starting in the late 1960s and accelerating in the late 1970’s when policy loan interest rates in most life insurance contracts were in the range of 5% to 6%, while market interest rates climbed to unprecedented levels. Bartlett, Tr. 3429-30, G736 at 29; McGill, Tr. 3697-3701, R163/G737 at 19-20; Donald J. Puglisi (“Puglisi”), Tr. 4343-45. When the prime rate reached 20% and the interest rate on six-month treasury bills reached more than 15%, G5 at 537, droves of policyholders made policy loans at low policy loan interest rates in order to invest in the market at higher interest rates. Bartlett, Tr. 3429-31, G736 at 29; McGill, Tr. 3700-01, R163/G737 at 9, 20-21. This policy loan drain threatened the solvency of some life insurance companies and placed others in a hazardous financial position, especially if they had to sell their long term, relatively low yielding investments at distress prices to raise cash to meet the burgeoning policy loan demand. Bartlett, Tr. 3430, G736 at 29; McGill, Tr. 3700-01, R163/G737 at 20-21; G5; Eisenberg, Tr. 2374-75. The differential crediting rates were created to curb the effects of disin-termediation. 6 Bartlett, Tr. 3430; McGill, Tr. 3697-3703; MCGILL’S LIFE INSURANCE at 84-87.

B. Development of the MBL COLI VIII Policy and Plan

The term “COLI” is an acronym which refers to “corporate owned life insurance” policies purchased by a corporate employer covering the lives of its officers or employees. Stip. ¶ 11. The corporate employer is the policyholder, premium payer, and beneficiary of a COLI policy. Id. The concept of COLI has existed for many years to help insure against the loss of key employees and to fund executive deferred compensation, supplemental retirement, and other benefit plans. MCGILL’S LIFE INSURANCE at 16, 18; McGill, G737 at 6. Common types of COLI policies, some of which were owned by Camelot, are known as “key man” policies, split dollar policies, and supplemental employee retirement policies (“SERP”). MCGILL’S LIFE INSURANCE at 314^15; Stip. ¶ 34.

The idea behind what has evolved into the MBL COLI VIII policy and plan was originally conceived in 1985 by Henry F. McCamish (“McCamish”), a life insurance entrepreneur. Eisenberg, Tr. 2163-65; 2168-69. McCamish retained Milliman & Robertson (“M & R”), a nationally prominent actuarial consulting firm, to perform the developmental and actuarial work for a new type of COLI product. Stip. ¶¶ 80-81; Eisenberg, Tr. 2163-65, 2168-69; McCamish, Tr. 4782-83. MeCamish’s instructions were 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 and that generated positive earnings and cash flow to the corporation in the first plan year. Stip. ¶ 81; Eisenberg, Tr. 2164-66; McCamish, Tr. 4782-85; G542 at Bates NJ3429. Eisenberg, managing consulting actuary of M & R’s life insurance practice, and Timothy Millwood, an actuary who joined M & R in 1987, worked on the project. 7 Stip. ¶ 81.

In addition, M & R was in charge of ensuring the new COLI policy and plan complied with all pertinent then-existing provisions of the I.R.C., including I.R.C. §§ 264 and 7702, 8 26 U.S.C. §§ 264, 7702. Eisenberg, Tr. 2169-70. I.R.C. § 264(a)(3) generally disallows interest deductions on amounts systematically borrowed on an insurance policy to pay premiums unless “no part of 4 of the annual premiums due during the 7 year period ... is paid under such plan by means of indebtedness.” 26 U.S.C. § 264(c)(1) (now codified at I.R.C. § 264(d)(1) but referred to herein as I.R.C. § 264(c)(1)) (known as the “four out of seven safe harbor”). Thus, M & R had to design a plan in which the policyholder would not take policy loans in four out of the first seven policy years.

MBL agreed to underwrite the new COLI product being designed by M & R. Stip. ¶ 81; Wendell J. Bossen (“Bossen”), Tr. 1733-35; James Van Etten (“Van Et-ten”), Tr. 2683-84, 2687-91; G22 at Bates MBL3271; G944 at Bates ICMG 249. As part of this arrangement, M & R performed the actuarial design for the COLI product under MBL’s general oversight. Id.

McCamish formed two corporations that provided administrative and consulting services for the new MBL COLI policies, Integrated Administration Services, Inc. (“IAS”) and IAS Development Corporation (“IDC”). Stip. ¶ 82; Bossen, Tr.1998; G22 at Bates MBL 3271; G944 at Bates ICMG 249. As compensation for these services, MBL agreed to pay McCamish and/or IAS and IDC a flat policy administration fee for every COLI policy sold by MBL, as well as a percentage of the annual premiums generated by the policies. Van Etten, Tr. 2691-93; McCamish, Tr. at 4787-91.

By early 1986, M & R, MBL, and McCamish developed policy form FA85, and supporting computer software, which was designated by MBL as “COLI I.” Eisenberg, Tr. 2180-82; G542 at 2-3. From 1986 to 1991, policy form FA85 was modified at various times in response to tax law changes and perceived competitive pressures in the marketplace. The modifications consisted of adding a series of “endorsements” or amendments to the form. Stip. ¶ 83. MBL designated this series of COLI products that were offered in the marketplace as COLI III, COLI V, COLI VIII, and COLI X. Id.

The COLI III policy was created in late 1986 in response to an amendment to I.R.C. § 264, which limited the deductibility of policy loan interest to interest to $50,000 per insured. Eisenberg, Tr. 2181-82; G542 at 3. The COLI III policy initiated the use of dividends to offset payment of premiums in four out of the first seven policy years. Id.

. The COLI V policy modified COLI III by making it into a “limited-payment plan,” i.e., making premiums payable for only a defined number of years at the beginning of the policy. Since the policy was designed so that all but four out of the first seven premiums would be paid by policy loan, the COLI V policy contemplated that premiums would stop being paid once the policyholder reached the $50,000 loan limit of I.R.C. § 264. Eisenberg, Tr. 2183-85; G542 at 3-4. Once the premi- urns stopped, the policyholder would hold its policies under an extended term nonfor-feiture option. Id.

Finally, the COLI VIII policy modified COLI V by reducing the amount of premium per thousand dollars of death benefits. The level of life insurance premiums is measured by the “premium per thousand,” the amount of premium dollar charged by the insurance company for every $1,000 of death benefits payable upon the death of the insured. Eisenberg, Tr. 2209-10, 2432; Kristie Sayre, 9 Tr. 3946. In 1986, Congress enacted I.R.C. § 7702A, which placed a cap on the premium per thousand in life insurance contracts. 26 U.S.C. § 7702A. If the premium per thousand exceeds this cap, the contract is considered a “modified endowment contract” and loses the tax benefits of a life insurance contract. Id. The COLI VIII policy reduced the premium per thousand in order to ensure compliance with I.R.C. § 7702A. 10 Eisenberg, Tr. 2185-86; Bossen, Tr.2004; G542 at 4.

C. Camelot’s Decision to Purchase the MBL COLI VIII Plan

Attracting and retaining quality employees was considered by Camelot’s owner, Paul David, to be critical to the success of the business. See Rl; Rogers, Tr. 153; Thomas Knoll (“Knoll”), Tr. 736. In order to accomplish this objective, Camelot developed the “Camelot Associate Program,” an employee benefit program offered to all employees who worked a minimum of 20 hours per week that included health and dental coverage in addition to other benefits. See J172; Rogers, Tr. 152, 155, 162-63. In the late 1980s, similar to other employers, Camelot experienced escalating health care costs, which were of mounting concern to the company. See J6; Jll; R325-28; Rogers, Tr. 159, 161, 165-66, 188; Knoll, Tr. 737; Barbara Bald (“Bald”), Tr. 855-56. Compounding the impact of the health care cost inflation of the late 1980s, the number of Camelot employees eligible for health benefits more than doubled, from 600 to 1300 associates, between 1987 and 1989. See J6; Stip. ¶ 25.

As part of his investigation into ways to finance the company’s increasing medical benefits costs, Rogers, Camelot’s Chief Financial Officer, spoke with Campisi of The Newport Group, Inc. (“Newport”) regarding possible funding vehicles. Rogers, Tr. 150-52, 169, 174; Campisi, Tr. 118-20. 11 Camelot was first introduced to the MBL COLI product as a possible means to fund employee health benefits in a letter from Campisi to Rogers dated July 6, 1987. J7/G934. In this letter, Campisi described the MBL COLI product and emphasized that “[t]he key factor is being able to absorb the interest deductions.” J7/G934. The July 6, 1987 letter, in an attached example, also illustrated to Rogers how higher policy loan interest rates advantaged the policyholder under the MBL COLI product. J7/G934, Bates 000071; Rogers, Tr. 408-09; Campisi, Tr. 1125-26, 1364-68.

Between 1987 and 1990, Rogers and Campisi met several times to discuss Camelot’s medical benefits cost funding issues. Rogers, Tr. 168-70, 174-76; Campisi, Tr. 1162-63. In addition Campisi sent a letter to Rogers, dated December 5, 1989, addressing a question Rogers had raised regarding the policyholder’s ability to terminate a COLI VIII plan. J15. Campisi stated that plan termination would only be desirable if the policy loan interest deduction could not be used by the corporation or if it was eliminated by Congress. J15, Bates 000461 & 000463; see also Rogers, Tr. 423-24, 427; Campisi, Tr. 1405-12.

Sometime between December 5 and December 22, 1989, Campisi sent an undated letter to Rogers enclosing a set of 40-year sales illustrations showing the projected cash flows and earnings performance of a COLI VIII plan based on a annual premium of $12,000 per employee and 1,294 covered employees. J188. The illustrations reflected premiums would be paid by policy loans in years one through three and through a combination of dividends and partial withdrawals in policy years four through seven. Campisi subsequently sent Rogers additional sales illustrations showing the projected financial performance of a COLI VIII plan with varying levels of premiums and number of covered employees. Based on these various sales illustrations, Rogers understood that Camelot could increase its aggregate policy loan, policy loan interest deduction, and after-tax cash flow by selecting a high level of premium. Rogers, Tr. 447-49, 450-51, 487; see also Campisi, Tr. 1441-43.

On December 22, 1989, Campisi sent a letter to Rogers enclosing two sets of 40-year sales illustrations showing the projected financial performance of a COLI VIII plan with an annual premium of $16,428 per employee and 1,294 covered employees, the only difference between the two sets of illustrations being that the second set assumed that no employees would die in the first plan year. J17; see also Rogers, Tr. 447-49; Campisi, Tr. 1464-66. The illustrations were based on payment of premiums by policy loans in years one through three and through a combination of dividends and cash value withdrawals in policy years four through seven. At some point in late 1989, Rogers asked Lee Ann Thorn (“Thorn”), Camelot’s Director of Tax, to project the company’s future taxable income to ensure that Camelot would be able to fully deduct the anticipated policy loan interest under the COLI VIII plan and thereby obtain the illustrated after-tax cash flows. Thorn, Tr. 668-69, 696, 699.

On February 13, 1990, Campisi forwarded to Rogers a binder package consisting of an application for a COLI VIII plan and a Prepayment Agreement. Campisi instructed Rogers to sign and return the enclosed forms along with a check in the amount of $200 times the number of employees to be insured under the Prepayment Agreement, which would provide interim coverage for a period of 60 days. Campisi’s letter explained to Rogers that the 60-day prepayment period “will allow us to tailor the size of the program to best fit Camelot’s taxable income expectations.” J22/J23. Campisi advised Rogers to date the forms “no later than February 20, 1990,” so that the policies, once issued, could be backdated to the application date, thereby increasing the probability that Camelot’s COLI VIII plan would be grandfathered in the event adverse legislation then being considered by Congress was enacted. J22/J23; see also Rogers, Tr. 484-87; Campisi, Tr. 1195-97.

In response, by letter dated February 16, 1990, Rogers forwarded to Campisi a signed application and Prepayment Agreement and a check in the amount of $278,200 to bind coverage and obtain interim death benefit coverage under the Prepayment Agreement for 60 days. J27; J28. Rogers indicated the 60 days would allow Camelot to evaluate the COLI VIII plan and also stated that Camelot wanted the ability to rescind its purchase if negative tax legislation were enacted at any time in 1990. J26 at Bates CAM004637-38; see also Stip. ¶ 55; Rogers, Tr. 236, 243^4, 493-94, 496; Campisi, Tr. 1197-98, 1496; J27; J28. At this point, Camelot had not yet determined the level of the annual premium or the number of insured employees under its COLI VIII plan. Rogers, Tr. 236, 243-44; G374, Bates 000079.

On February 16, 1990, Rogers wrote a file memorandum titled “COLI Summary” that documented various issues related to Camelot’s COLI VIII purchase, including the company’s efforts to minimize the risk that the policy loan interest deduction would be eliminated by legislation and the potential impact to Camelot of “freezing” the COLI VIII plan if adverse tax legislation were enacted or Camelot could not absorb the interest deductions. J25. Rogers’ memorandum concluded: “[t]he decision to go forward and institute a COLI plan boil down to the risks involved. Those risks specifically include: 1) A retroactive tax law change[,] 2) Camelot’s failure to generate taxable income over several years in a row[, and] 3) IRS attackf.]” J25; see also Rogers, Tr. 245-47, 501, 507.

During the prepayment period, Camelot evaluated its potential COLI VIII purchase, including soliciting advice from its outside tax advisers, Ernst & Young, and its outside legal counsel, Stark & Knoll. During this period, Camelot also evaluated various options regarding the premium level and number of insureds. By letter dated March 26, 1990, Campisi forwarded to Rogers two sets of sales illustrations projecting the financial performance of a COLI VIII plan assuming an annual premium per employee of $16,667 and $12,000 (characterized by Campisi as “Scenario I” and “Scenario II,” respectively), 1,432 insured employees, and for payment of premiums by policy loans in the first three years and by dividends and cash withdrawal in years four through seven. J33. Cam-pisi recommended Camelot choose the higher premium if Camelot could take advantage of the larger tax deductions: “[I]f you feel comfortable with Camelot’s ability to absorb the higher tax benefit during the first seven years, Scenario I [the higher annual premium per employee of $16,667] would be an appropriate choice.” J33; see also Rogers, Tr. 299-300, 509-11; Campi-si, Tr. 1497-98.

On April 11, 1990, Camelot held a meeting at its offices to finalize the design of Camelot’s COLI VIII plan. Campisi gave a presentation and distributed an “Executive Summary” describing the COLI VIII plan as a “financial tax-advantaged strategy” that would generate positive cash flows in every plan year without materially competing for Camelot’s use of capital. R406; see also J34; Rogers, Tr. 252-254; Campisi, Tr. 1200-02.

By letter dated April 24, 1990, Campisi sent Rogers three additional sets of sales illustrations projecting the financial performance of the COLI VIII plan based on (1) a premium of $16,667 on 1,000 employees, (2) a premium of $10,000 on 1,432 employees, and (3) a premium of $10,000 on 1,432 employees with the tax savings from the policy loan interest deduction artificially capped at $2.5 million. J35/G403. In the letter, Campisi recommended Camelot choose a premium of $10,000 per employee. J35/G403; Rogers, Tr. 519-23; Campisi, Tr. 1213-14, 1501-02. Camelot followed Campisi’s advice and elected an annual premium of $10,000 on 1,431 employees, 12 covering all employees working at least 20 hours a week, ranging from Chairman and owner Paul David to part-time, entry-level sales associates at Camelot stores. Stip. ¶¶ 56, 61, 72; Rogers, Tr. 297-98, 314-16, 419-20. On April 30, 1990, Camelot issued a check payable to MBL in the amount of $707,719.63, the balance owing on the first-year premium. Stip. ¶ 64; G405; J39.

By memorandum dated May 1, 1990, McCamish’s IAS forwarded to Newport three sets of 81-year illustrations for Camelot’s COLI VIII plan. J42. Two of the sets of illustrations generated by IAS, labeled “Scenario 1” and “Scenario 2,” projected the financial performance of Camelot’s plan assuming that Camelot paid the annual premiums with cash or financing external to the policy in all or some of the first seven plan years. J42. “Scenario 3” projected the financial performance of the plan assuming payment of premiums predominantly through a combination of policy loans, dividends and partial withdrawals, similar to the prior sales illustrations provided to Camelot. Rogers did not recall seeing these illustrations prior to this litigation. Rogers, Tr. 298-99, 465-66.

By letter dated May 3, 1990, Campisi forwarded to Rogers issue illustrations projecting the financial performance of Camelot’s COLI VIII plan for 20 years, reflecting that premiums would be offset with policy loans in the first three plan years and a combination of loading dividends and partial withdrawals in the fourth through seventh plan years. J44; see also Rogers, Tr. 536-37; Campisi, Tr. 1241-42. On May 21,1990, Newport delivered the COLI VIII plan policies to Rogers at Camelot’s offices in Ohio. Stip. ¶ 70; J45; Rogers, Tr. 548.

D. Features of Camelot’s COLI VIII Policies and Plan

The COLI VIII policies purchased by Camelot were issued by MBL on Policy Form FA85 with endorsements EFA85-3, EFA85-4a, EFA85-6, and EFA 85-7. Stip. ¶ 12. Camelot’s witnesses have characterized Camelot’s COLI VIII policies in varied ways. Each COLI VIII policy was denominated as an “Increasing Death Benefit Whole Life” policy with “Premiums Fixed and Payable During Lifetime of Insured or Until End of Premium Period.” J189 at Bates 001064; see also Stip. ¶ 15. Camelot’s actuarial expert, DesRochers, characterized the COLI VIII policies as “increasing [death] benefit whole life contracts” with “fixed premiums,” which use “universal life type mechanics to generate policy values.” DesRochers, Tr.1950. He testified that, within the industry, a COLI VIII policy would probably be characterized as an interest-sensitive whole life policy or a fixed-premium universal life policy. Id. The chief actuarial architect of the COLI VIII plan, Eisenberg, has described the COLI VIII policy variously as a fixed-premium universal life policy, Eisenberg, Tr. 2189, a fixed-premium increasing death benefit whole life policy, id. at 2578, and an interest-sensitive participating whole life policy which, from a risk perspective, is annual, renewable term insurance. Id. at 2582.

Rather than be confused by the varying characterizations of the COLI VIII policies, the operation of the complex COLI VIII policies and plan 13 can best be understood by describing their features and function. To understand the mechanics of the COLI VIII policies, the parties have stipulated to the use of a sample COLI VIII policy covering a 28 year old male. J189. This policy is representative of the 1,430 COLI VIII policies Camelot purchased. In addition, to understand the mechanics of the highly structured, leveraged financing plan for the purchase of the COLI VIII policies, it is helpful to consider some of the “Issue Illustrations” Newport provided to Camelot.

Appendix A contains a modified version of the illustration, J44, which Camelot received immediately after it purchased the COLI VIII plan. J44 illustrates the projected cash flow of Camelot’s COLI VIII plan over 20 years. Columns (2), (3), and (4) illustrate the premium, administrative fee, and accrued policy loan interest that was projected to be due at the beginning of each policy year. Columns (5), (6), (7), and (8) show the aggregate amounts of policy loans, loading dividends, partial withdrawals, and cash used to pay the amounts due at the beginning of each policy year. Column (9) reveals the net equity remaining in the policy at the end of a given policy year. Column (10) displays the amount of net death benefits Camelot was projected to receive during the course of a given policy year. Column (11) illustrates the projected annual cash flow before taking into account the policy loan interest deductions. Column (12) shows the amount of the tax benefit from the interest deductions in each policy year. Finally, column (18) displays the net annual cash flow after taking into account the interest deductions in each policy year. Appendix B contains a similar modified illustration showing the actual values of each of these features for Camelot’s COLI VIII plan over the first seven policy years. The mechanics of the COLI VIII plan can best be understood by elaborating on each of these features.

The COLI VIII policies were designed to have fixed annual premiums (column (2)), payable on the first day of each policy year. The policyholder could select the magnitude of the annual premiums, ranging from $2,000 to $16,667 per policy, to tailor the plan to its needs. Camelot selected an annual premium of $10,000 per employee. The total annual premium due on the first day of the first policy year was equal to $10,000 multiplied by 1,430 COLI VIII policies, or $14,300,000. 14 As employees died, their policies would cease to be in effect and Camelot would no longer pay premiums on those policies. Thus, in each subsequent year, the total annual premium would be equal to the number of employees still alive multiplied by $10,000. In addition, Camelot planned to stop paying premiums after the ninth policy year and have the policies go into “paid-up status.” 15

Finally, because Camelot was projected to take a policy loan to pay the premium, it had to pay accrued interest on the policy loan. The accrued interest was payable on the first day of the following policy year. For example, interest on a policy loan taken on the first day of the first policy year was due on the first day of the second policy year. Column (4) reflects the projected accrued policy loan interest due on the first day of each policy year.

2. Sources of Funds to Pay Amounts Due (Columns (5), (6), (7), and (8))

a. Policy Years One through Three

In the first through third and eighth and ninth policy years, Camelot’s COLI VIII plan was designed for Camelot to pay approximately 90% of the annual premiums and accrued policy loan interest through a policy loan, and the remaining 10% in cash, in a simultaneous netting transaction. 16 On the first day of the first policy year, the following transactions occurred simultaneously: 17 (i) Camelot paid a premium of approximately $14 million to MBL, creating a policy value of approximately $14 million; (ii) Camelot took a policy loan of approximately $13 million from MBL, using the policy value as collateral; (iii) the $13 million policy loan was used to offset payment of the $14 million premium; and (iv) Camelot paid approximately $1 million in cash to MBL. 18 Stip ¶¶ 64, 97, 167. On the first day of the second and third policy years, similar simultaneous netting transactions occurred, whereby: (i) Camelot paid the premium plus accrued loan interest; (ii) Camelot took a policy loan equal to approximately 90% of the annual premium plus accrued loan interest; (iii) the policy loan was used to offset payment of the premium plus accrued loan interest; and (iv) Camelot paid the approximately 10% balance by cash. 19 Stip. ¶¶ 100, 103, 149, 153, 169, 171; J54; J71; Kristie Sayre, Tr. 4017-4020, 4036; R314/G738 at 42. These simultaneous netting transactions were necessary to create the policy value needed as collateral to support making the policy loan. Eisenberg, Tr. 2418-25; Kristie Sayre, Tr. 4016-17, 4019-20, 4036; R314/G738 at 42-46. Because of the broad base of employees and the pattern of borrowing, the COLI VIII plan is sometimes referred to as “broad-based leveraged COLI.” See, e.g., Tr. 7, 8, 14, 44, 68, 89, 91.

b. Policy Years Four through Seven

In policy years four through seven, the annual premiums and accrued loan interest were designed to be paid by a combination of cash, a loading dividend, and a partial withdrawal, in another simultaneous netting transaction. On the first day of each of the fourth through seventh policy years the following simultaneous transactions were deemed to have occurred: (i) Camelot paid the annual premium plus accrued interest; (ii) approximately 95% of the annual premium was taken by MBL or Hartford as an expense charge, while approximately 5% was credited to the policy value; (iii) approximately 5-8% of the expense charge was set aside to cover MBL or Hartford’s actual expenses; (iv) approximately 92-95% of the expense charge was immediately returned to Camelot in the form of a “loading dividend”; (v) Camelot received a partial withdrawal of policy value in an amount equal to approximately 99% of the accrued loan interest; (vi) the loading dividend and partial withdrawal were used to offset payment of the annual premium and accrued loan interest; and (vii) Camelot paid the balance due in cash. Eisenberg, Tr. 2453-54, 2583; Bartlett, Tr. 3398-3402, G736 at 5-6, 19-20; Kristie Sayre, Tr. 4036-38, 4042^16, 4056, 4059-60, R314/G738 at 42-43, 47-48; J93; J109; J115; J123; J128; G738 at 40; G527 at Bates 027108-09.

The loading dividends are sourced from the expense charge, also called the “loading charge.” Each Camelot COLI VIII policy contractually provides the magnitude of the loading charge (called “premium expense charge rate”) as a percentage of the annual premium. J189 at Bates 001135. The expense charge covers the anticipated actual expenses plus a margin for a loading dividend. The designers of the COLI VIII policies issued actuarial memoranda with the anticipated expenses for each year. See, e.g., G877. The following chart compares the loading charges and anticipated expenses for each year, expressed in terms of percentage of premium, for a 25 year-old male: 20

Anticipated Year Loading Charge Expenses Difference

1_10,00%_2,00%_8.00%

2_20,00%_6.80% 13.20%

3_30.00%_6.80% 23.20%

4-7_95.00%_7.30% 87.70%

8-9_70,00%7.30% 62.70%

See G887. Because the loading charge is projected to exceed the anticipated expenses by approximately 87.7% in policy years four through seven, the COLI VIII design proposed that this excess be returned to Camelot in the form of a loading dividend. Bartlett, Tr. 3399; G736 at 19-20; G844 at Bates 5583. The amount of the loading dividend is equal to the difference between the expense charge and the actual expenses for the COLI VIII policies. McGill, R1631/G737 at 53.

The partial withdrawals are deducted from the policy value of each COLI VIII policy. The policy value is generally equal to the total annual premiums, less expense charges, plus the inside build-up on the annual premium, plus any dividend credited to policy value, less the COI charges. J189 at Bates 001073. The COLI VIII policy provides Camelot may take a partial withdrawal of the policy value. J189 at Bates 001093 (Endorsement EFA 85-7); Bossen, Tr. 1743; DesRochers, Tr. 4859. A partial withdrawal reduces policy value and death benefits. Eisenberg, Tr. 2495; DesRochers, Tr.1983. The COLI VIII policy places no limit on the amount of policy value that may be taken as a partial withdrawal. Bossen, Tr. 1758.

However, if the policy has zero net equity, any partial withdrawal may only be used to pay off policy loans and accrued policy loan interest, and not to pay the premium. The policy states “[w]e will use the amount withdrawn to reduce the loan balance to the new loan limit, and then we will pay the rest in cash.” J189 at Bates 001093. The “loan balance” is the loan principal plus accrued interest less unearned interest. J189 at Bates 001069. The “loan balance” cannot exceed the “loan limit,” which, on the last day of a policy year, is the policy value plus the dividend value. J189 at Bates 001090 (Endorsement EFA 85-4a). “Cash surrender value,” or net equity, is equal to the policy value plus dividend value less the policy loan and accrued interest, i.e., the loan balance. J189 at Bates 001089 (Endorsement EFA 85-4a). In other words, on the last day of a policy year, the net equity equals the loan limit less the loan balance. If the net equity in the policy is zero, i.e., the loan balance is equal to the loan limit, a partial withdrawal will be fully used to lower the loan balance on a dollar for dollar basis.

For example, assume the policy has a policy value of $10,000 and there is an outstanding loan balance of $10,000, consisting of $8,500 in loan principal and $1,500 in accrued interest. Although the policy has a net equity of zero, the policyholder may make a partial withdrawal of an amount up to the policy value of $10,-000. If the policyholder makes a partial withdrawal of $2,000, the policy value will decrease to $8,000.

3. Loan Interest Rate and Differential Crediting Rates

Like traditional universal life policies, Camelot’s COLI VIII policy had three interest rates: the policy loan interest rate, the loaned crediting rate (called the “Current Credited Loaned Interest Rate” in Camelot’s COLI VIII plan) and the un-loaned crediting rate (called the “Current Credited Unloaned Interest Rate” in Camelot’s plan). The unloaned crediting rate was always the greater of a rate declared by MBL or 4%. J189 at Bates 001089.

The loan interest rate and the loaned crediting rate were set as follows. The COLI VIII policies permitted Camelot to annually elect between a fixed or variable loan interest rate. J189 at Bates 001077-80. The fixed loan interest rate was 8% per annum if charged in arrears or 7.4% if charged in advance. J189 at Bates 001080. If Camelot elected the fixed loan interest rate, the loaned crediting rate would be set by MBL but would be no less than 4%. J189 at Bates 001080.

If Camelot elected the variable loan interest rate, the loan interest rate would be the greater of: (1) the Moody’s Corporate Bond Yield Average — Monthly Average Corporate (hereinafter “Moody’s Corporate Average”) or (2) the loaned crediting rate plus 1%. J189 at Bates 001078. Under Endorsement EFA85-3, the loaned crediting rate was calculated according to the mathematical formula (hereinafter “Baa enhanced”):

Baa + T%

100% — Baa

where Baa refers to Moody’s Corporate Bond Yield Average — Monthly Corporate Baa (“Moody’s Baa”). Camelot could elect between T=0% and T=l%. J189 at Bates 001088; Eisenberg, Tr. 2256-59; McGill, Tr. 3722-23, R163/G737 at 23. 21

Moody’s Baa has historically exceeded Moody’s Corporate Average by about 0.60% to 1.00%. Eisenberg, Tr. 2291-94, 2323-24; McGill, Tr. 3720-21, R163/G737 at 23; G935. Mathematically, the Baa enhanced loaned crediting rate will always yield a rate which is greater than Moody’s Baa. Eisenberg, Tr. 2258-59, 2326-27; McGill, Tr. 3722-23, R163/G737 at 23. Therefore, the Baa enhanced loaned crediting rate is also always greater than the Moody’s Corporate Average. Accordingly, the second prong of the variable loan interest rate provision will always govern. Eisenberg, Tr. 2326-27.

To summarize, if Camelot elected the variable loan interest rate, it could elect between two Baa enhanced loaned crediting rates, and the loan interest rate would automatically be 1% greater than the loaned crediting rate. Eisenberg, Tr. 2324-26; Campisi, Tr. 1332-33; Kristie Sayre, Tr. 3964-65, R314/G738 at 32; G56 at Bates NJ650; G958. The following chart illustrates the three choices of loan interest rate and loaned crediting rate combinations available to Camelot under its COLI VIII policy:

Loaned Crediting Rate Loan Interest Rate

Set by MBL or Fixed Rate of 4% Fixed Rate of 8%

Baa + 0% Baa + 0%

1 — Baa 1 — Baa +1%

Baa + 1% Baa + 1%

1 — Baa 1 — Baa +1%

Campisi, Tr. 1332-33; Kristie Sayre, Tr. 3964-65; R314/G738 at 32; G958. In each relevant policy year, Camelot elected the variable loan interest rate option. Camelot also selected the Baa enhanced loaned crediting rate with T=l%, which generated the highest possible loan interest rate. Rogers, Tr. 291-93, 396; Hoag, Tr. 4578-80.

4. Net Equity (Column (9))

The COLI VIII policy was 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 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 said to be encumbered and the policy value cannot support taking an additional policy loan. The undisputed evidence shows the policy achieved a net equity equal to zero on the last day of each policy year. Eisenberg, Tr. 2226, 2518-22; Kristie Sayre, Tr. 3994-96, 4002-09, 4010-11; R314/G738 at 40; G792; G844 at Bates 5583; G848 at 2; G949 at Bates 015150; G972; G987.

5. Death Benefit (Column (10))

Camelot’s COLI VIII policy was an increasing death benefit policy, meaning that the death benefit could grow to exceed the original face amount of the policy. Each COLI VIII policy had a different face amount based upon the age and gender of the insured. According to the sample COLI VIII policy:

The death benefit on any date is the greater of:

(a) the specified [face] amount ... and

(b) the policy value on that date divided by the net single premium based on the sex, age, and rate class of the insured on that date....

J189 at Bates 001071-001072. Thus, as the policy value increased, the death benefit could also increase over the specified face amount.

The death benefit shown in Appendix A is the projected net death benefit, which is the amount of proceeds Camelot was projected to receive upon the death of insured employees in any particular year. The net death benefit is the sum of:

(d) the death benefit;

(e) the amount of any dividend additions;

(f) any amount payable under an extra benefit rider; and

(g) if the insured dies during a period which is covered by a payment of premiums, a refund of that part of any premium which did not increase the death benefit based on the number of days between the date of death and the end of the period;

reduced by:

(h) any loan balance; and

(i) if death occurs during the grace period, any unpaid premiums.

J189 at Bates 001072.

6.Net Pre-Interest Deduction Cash Flow (Column (11))

Column (11) shows the net cash flow Camelot was projected to receive absent the tax benefit from the policy loan interest deductions. This number is simply equal to the net death benefit (Column 10) less the cash Camelot paid out of pocket (Column (8)).

7. Interest Deduction Tax Benefit (Column (12))

Camelot was projected to receive each year a tax benefit from a deduction for the interest paid on the policy loan, pursuant to 26 U.S.C. § 163. The tax benefit was to be offset from Camelot’s federal taxable income in an amount equal to the policy loan interest paid that year. Because Camelot was projected to have a corporate income tax rate of 40%, the tax benefit of the interest deduction would be 40% of the amount of the deduction. It is these interest deductions which are the subject of the dispute in this ease.

8. Net After Interest Deduction Cash Flow (Column (13))

Column (13) shows the net cash flow Camelot was projected to receive after taking into account the tax benefit from the interest deductions. This is computed by summing the pre-interest deduction cash flow (Column (11)) and the interest deduction tax benefit (column (12)).

One word best captures the complex, highly-calibrated COLI VIII plan that Camelot purchased — “innovative.” Eisen-berg, Tr. 2671-75. The COLI VIII plan contains an innovative combination of features, some of which are innovative in their own right, designed to generate positive cash flows every policy year. First, the COLI VIII plan had a very high policy value on the first day of the first year of the policy. The high policy value was made possible by high premiums per employee and in the aggregate because the plan included purchase of policies on a broad-base of 1,430 employees, as well as agents forgoing first year commissions. Second, the high policy value was used to secure maximum policy loans, which paid the high premiums in the first three policy years and sourced substantial interest deductions that were intended to fuel the positive cash flows. Third, computer technology enabled maximum leveraging of these high policy values through programs designed to achieve zero net equity at the end of each policy year. Fourth, Camelot had the right to determine among crediting rates for the inside build-up on the loaned portion (i.e., virtually all) of the policy value, which necessarily determined the policy loan interest rate because of the 1% spread between the loaned crediting rate and policy loan rate. Fifth, the 1% spread essentially fixed Camelot’s cost of borrowing with the counterintuitive result that the higher the loaned crediting interest rate paid by Camelot, the greater the cash flows to it as a result of higher interest deductions. Sixth, the COLI VIII plan had an extremely high expense loading component in policy years four through seven, which expense loads were used to create first day dividends 22 intended to pay a substantial portion of the premiums.

In sum, through a pre-planned, highly-structured and calibrated combination of innovative features, Camelot’s broad-based leveraged COLI VIII plan was designed to produce positive cash flows in each and every plan year, assuming the deductibility of the policy loan interest. The question which the Court must answer is whether these innovative features, either singly or in combination, crossed the line so as to render the policy loan interest non-deductible.

III. DISCUSSION

The IRS challenges Camelot’s policy loan interest deductions on three alternative bases. First, the IRS asserts the policy loan interest was not deductible under I.R.C. § 163 because the COLI VIII policy and plan were sham transactions. Second, the IRS contends Camelot’s COLI VIII policy and plan comprise a generic tax shelter. Third, the IRS argues the policy loan interest deductions should be disallowed under I.R.C. § 264(a)(3) because Camelot failed to pay four out of the first seven annual premiums due by a means other than indebtedness in order to qualify for the safe harbor of I.R.C. § 264(c)(1). In addition, the IRS asserts Camelot is liable for accuracy related penalties for substantial understatement of its income tax under I.R.C. §§ 6662(a) and 6662(b)(2). These four issues will be addressed in turn.

A. Deductibility of Interest Under § 163 and the Sham Transaction Doctrine

1. General Principles

Section 163(a) of the I.R.C. provides “there shall be allowed as a deduction all interest paid or accrued within the taxable year on indebtedness.” 26 U.S.C. § 163(a). At issue is whether the amounts characterized as policy loan interest under Camelot’s COLI VIII plan were, in fact, “interest” on “indebtedness” for federal tax purposes.

It has been long established that, for federal income tax purposes, interest is defined as “compensation for the use or forbearance of money.” Deputy v. du Pont, 308 U.S. 488, 498, 60 S.Ct. 363, 84 L.Ed. 416 (1940); see also Old Colony R.R. Co. v. Comm’r, 284 U.S. 552, 560, 52 S.Ct. 211, 76 L.Ed. 484 (1932); Weller v. Comm’r, 270 F.2d 294, 296 (3d Cir.1959), cert. denied, 364 U.S. 908, 81 S.Ct. 269, 5 L.Ed.2d 223 (1960). Interest can be deducted only if paid with respect to actual or genuine indebtedness. See Knetsch v. United States, 364 U.S. 361, 369, 81 S.Ct. 132, 5 L.Ed.2d 128 (1960); Goldberg v. United States, 789 F.2d 1341 (9th Cir.1986); Bridges v. Comm’r, 325 F.2d 180 (4th Cir.1963). Genuine indebtedness does not exist under § 163 if the underlying transaction is a sham transaction aimed solely at tax avoidance. See United States v. Wexler, 31 F.3d 117, 122 (3d Cir.1994), cert. denied, 513 U.S. 1190, 115 S.Ct. 1251, 131 L.Ed.2d 133 (1995); Weller, 270 F.2d at 297; CM Holdings I, 221 B.R. at 722.

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 in denying deductions claimed by taxpayers for losses and expenses incurred in a corporate reorganization. Although the taxpayers had followed each step required by the I.R.C. for the reorganization, the Court nonetheless held these losses nondeductible, reasoning 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, as opposed to formal reality, it was not “the thing which the statute intended.” Id.

The sham transaction doctrine requires a thorough examination by courts of the challenged transaction as a whole, as well as each step thereof, to determine if the substance of the transaction is consistent with its form. See ACM Partnership v. Comm’r, 157 F.3d 231, 246 (3d Cir.1998), cert. denied, 526 U.S. 1017, 119 S.Ct. 1251, 143 L.Ed.2d 348 (1999); Weller, 270 F.2d at 294. If a transaction’s form complies with the Code’s requirements for deductibility, but the transaction lacks the factual or economic substance that the form represents, then expenses or losses incurred in connection with the transaction are not deductible. See Knetsch, 364 U.S. at 365-66, 81 S.Ct. 132; Wexler, 31 F.3d at 122; Lerman v. Comm’r, 939 F.2d 44, 45 (3d Cir.1991), cert. denied, 502 U.S. 984, 112 S.Ct. 590, 116 L.Ed.2d 615 (1991); Kirchman v. Comm’r, 862 F.2d 1486, 1490 (11th Cir.1989). “To permit the true nature of a transaction to be disguised by mere formalisms, which exist solely to alter tax liabilities, would seriously impair the effective administration of the tax policies of Congress.” Commissioner v. Court Holding Co., 324 U.S. 331, 334, 65 S.Ct. 707, 89 L.Ed. 981 (1945).

Courts have recognized two basic types of sham transactions: shams in fact and shams in substance. See ACM, 157 F.3d at 247 n. 30 (citing Kirchman, 862 F.2d at 1492). “[Sjhams in fact” are transactions that never occurred in reality, that is, transactions that have been created on paper but which never took place. See id. “[Sjhams in substance” are transactions that actually occurred but are devoid of economic substance. See id.; accord Lerman, 939 F.2d at 49 n. 6.

The IRS contends the evidence adduced at trial demonstrates that both the sham in fact and the sham in substance doctrines apply to Camelot’s COLI VIII plan. It urges the Court to disallow the policy loan interest deductions under I.R.C. § 163(a). The burden is on the taxpayer, Camelot, to prove by a preponderance of the evidence that the form of the transaction reflects its substance. See Goldberg, 789 F.2d at 1343; accord National Starch and Chem. Corp. v. Comm’r, 918 F.2d 426, 429 (3d Cir.1990), aff'd sub nom, INDOPCO, Inc. v. Comm’r, 503 U.S. 79, 112 S.Ct. 1039, 117 L.Ed.2d 226 (1992). See also, e.g., New Colonial Ice Co. v. Helvering, 292 U.S. 435, 440, 54 S.Ct. 788, 78 L.Ed. 1348 (1934); Welch v. Helvering, 290 U.S. 111, 115, 54 S.Ct. 8, 78 L.Ed. 212 (1933). 23 The Court will discuss the sham in fact and sham in substance doctrines in turn after first addressing two preliminary arguments raised by Camelot.

Camelot devotes a considerable portion of its post-trial briefing to the argument that Camelot has complied with the detailed statutory regime regarding the de-ductibility of interest on life insurance policy loans as set forth in I.R.C. §§ 7702, 7702A, and 264. Camelot urges compliance with these provisions gives it the right to claim interest deductions and contends that if the Court were to disallow its policy loan interest deductions, it would effectively be legislating tax policy and usurping the power of Congress. The Court disagrees.

Adopting Camelot’s position would do away with the well-established sham transaction doctrine. Under the sham transaction doctrine compliance with the letter of the tax code does not confer a right to claim interest deductions if the transaction is a sham. See, e.g., Gregory v. Helvering, 293 U.S. 465, 468, 55 S.Ct. 266, 79 L.Ed. 596 (1935) (concluding the transaction lacked substance for tax purposes, even where the transactions on their face satisfied “every element required by” the relevant statutory language); ACM, 157 F.3d at 246 (“even where the ‘form of the taxpayer’s activities indisputably satisfies the literal requirements’ of the relevant statutory language, the courts must examine ‘whether the substance of those transactions was consistent with their form’ ” (citations omitted)). See also Winn-Dixie Stores, Inc. v. Comm’r, 113 T.C. 254, 290-94, 1999 WL 907566 (1999). The Court rejects Camelot’s argument. Instead it will conduct a probing inquiry into whether the substance of the transaction is reflected in its form.

In a similar vein, Camelot also argues the Court should not give retroactive effect to Congress’ decision in 1996 to eliminate COLI policy loan interest deductions prospectively through a phased-out “soft landing” when it in enacted the Health Insurance Portability and Accountability Act of 1996 (HIPA), Pub.L. No. 104-191, 110 Stat.1936, 2090. Like Camelot, the taxpayer in Winn-Dixie, 113 T.C. 254, 1999 WL 907566, argued the 1996 HIPA legislation demonstrated that its deductions for COLI policy loan interest were condoned by Congress. The Tax Court squarely rejected this argument, stating “we are not persuaded that Congress, by enacting and amending section 264 or other related provisions that restrict the deductibility of interest, intended to allow interest deductions under section 163 based on transactions that lacked either economic substance or business purpose.” Id. at 293. This Court agrees.

Contrary to Camelot’s position, Congress did not, by enacting the 1996 HIPA legislation, approve of interest deductions related to COLI VIII plans that lacked economic substance. As noted in a report by the Joint Committee on Taxation, “the proposal would not affect the determination of whether interest is deductible under present law rules (including whether interest paid or accrued during the phase-in period is otherwise deductible).” Description Of Revenue Provisions Contained In The President’s Fiscal Year 1997 Budget Proposal (Released On March 19, 1996), Staff of the Joint Committee on Taxation, at 82 (March 27, 1996) (emphasis added). In describing the present law rules regarding the deductibility of policy loan interest, the Conference Report stated that “[pjrovided the transaction gives rise to debt for Federal income tax purposes, and provided the 4-out-of-7 rule is met, a company may under present law borrow up to $50,000 per employee.... ” H.R. CONF. REP. NO. 104-736, at 319-20 (1996), reprinted in 1996 U.S.C.C.A.N. 1990, 2132-33. The Conference Report further noted that “[i]n addition to the specific disallowance rules of section 264, generally applicable principles of tax law apply,” id. at 320 n. 23, and “no inference, [was] intended as to the treatment of interest paid or accrued under present law.” Id. at 322. Stated another way by the Joint Committee report: “[T]he IRS would not be precluded from applying common-law doctrines or statutory or other tax rules to challenge corporate-owned life insurance plans to which present law rules apply.” Description Of Revenue Provisions Contained In The President’s Fiscal Year 1997 Budget Proposal (Released On March 19, 1996), Staff of the Joint Committee on Taxation, at 82 (March 27, 1996). Thus, the legislative history of HIPA makes clear that, in enacting restrictions on the deductibility of COLI policy loan interest, Congress did not intend to negate the continued validity of the sham transaction doctrine. 24

2. Sham in Fact

Factual shams are transactions which were created on paper but “[did] not occur in fact, or that ‘were performed in violation of some of the background assumptions of commercial dealing.’ ” Peerless Indus., Inc. v. U.S., 1994 WL 13837 at *4 (E.D.Pa.1994), aff'd, 37 F.3d 1488 (3d Cir.1994) (quoting Horn v. Comm’r, 968 F.2d 1229, 1236 n. 8 (D.C.Cir.1992)); see also ACM Partnership, 157 F.3d at 247 (citing Kirchman, 862 F.2d at 1492). Camelot must prove the purported transaction actually took place in a manner which did not deviate from relevant commercial norms. See Sheldon v. Comm’r, 94 T.C. 738, 753-58,1990 WL 69233 (1990); see also Forseth v. Comm’r, 85 T.C. 127, 165, 1985 WL 15375 (1985) 25 (finding purported commodity transaction to be factual sham, in part, because there was no evidence commodities actually traded); Krumhorn v. Comm’r, 103 T.C. 29, 39, 1994 WL 374421 (1994) (finding transaction to be factual sham, in part, because it “lacked common business formalities” and departed from industry norms). In evaluating whether a transaction constitutes a factual sham, the Court “will ignore accounting tricks and other transactional artifices.” Peerless, 1994 WL 13837 at *4.

Eisenberg repeatedly described features of the Camelot COLI VIII plan as “innovative.” In a factual sham analysis, the Court must determine whether these features, singly or together, are merely “innovative” variations of accepted life insurance policies, or are ingenious and contrived artifices that contravene accepted industry norms. The IRS asserts the following components of the Camelot COLI VIII plan are factual shams: the policy loans, the loading dividends, the partial withdrawals, and the annual premiums. The Court will evaluate each of these components in turn and then evaluate the COLI VIII plan as a whole.

a. Policy Loans

Camelot has proven by a preponderance of the evidence that these policy loans occurred in reality. Camelot was charged and paid significant interest on these policy loans. Stip. ¶¶ 167, 169, 171, 173-179; J189 ¶ 16. The policy loans also were considered real debts of Camelot, which had to be repaid at the time of death of the insured or when the policy lapsed. Bos-sen, Tr. 1722. When an insured Camelot employee died, a death benefit, reduced by the outstanding policy loan and unpaid interest, was paid to Camelot. Rogers, Tr. 326-29, 333-34, 336, 346-50; Campisi, Tr. 1252-1263; Eisenberg, Tr. 2267-68; Van Etten, Tr. 2796-99; J80; J82; J84; R52-R55; R58; R59; see also Coors v. United States, 215 Ct.Cl. 840, 572 F.2d 826, 834 & n. 14 (1978) (policy loans not illusory when repaid out of death benefits). Also, MBL and Hartford carried the policy loans as assets on their books and recorded them separately from the receipt of the premium. Bossen, Tr. 2128.

Despite this evidence, the IRS asserts the policy loans were factual shams because: (1) the loans were made on the first day of each of the first three policy years; (2) the loans lacked the risk characteristics of ordinary commercial and policy loans; (3) the loans had artificially high interest rates; (4) interest on the first-year loan was paid retroactively for a period preceding the date the loan was made; (5) MBL did not restrict COLI policy loans during rehabilitation; and (6) only 40% of accrued interest on the loans was paid in cash. These characteristics of the policy loans will be addressed in turn.

(1) Policy Loans Made on First Day of Policy Year

The IRS contends the policy loan violated background assumptions of commercial dealing because they were made on the first day of each policy year, at a time when there was zero net equity in the policy to support a policy loan. 26 As previously rehearsed, the undisputed evidence shows the policy had a net equity of zero immediately prior to the first day of any policy year. Eisenberg, Tr. 2225-26, 2518-22; Kristie Sayre, Tr. 3994-96, 4002-09, 4010-11; R314/G738 at 40; G792; G844 at Bates 5583; G848 at 2; G949 at Bates 015150; G972; G987. But, the preponderance of the evidence also indicates it is relatively common for insurance companies to make policy loans on the first day of any policy year, even when the policy has zero net equity.

The IRS primarily relies on the testimony of its actuarial expert, Kristie Sayre, who analogized the simultaneous netting transaction to the well-known riddle of the chicken and the egg, where it is impossible to tell which came first, the payment of the premium or the policy loan. Kristie Sayre, Tr. 4194. Although she could not point to any actuarial standards that require it, she testified policy loans are not real unless they are borrowed against policy value that has been in the policy for a defined period of time. Kristie Sayre, Tr. 4170-71, 4194-98. When asked how long policy value had to be in the policy before a policy loan could be taken, Kristie Sayre stated at least one, and preferably, five years. Kristie Sayre, Tr. 4194-98. She stated that a period of six months made her “uncomfortable” and that borrowing a policy loan on the first day was definitely not real. Kristie Sayre, Tr. 4194-98. Kristie Sayre also testified she knew of no other policy which allowed a first year, first day policy loan. Kristie Sayre, Tr. 4020.

Nonetheless, the IRS position and Kristie Sayre’s testimony flies in the face of the fundamental principle, accepted by both parties, that policy loans are made from the general funds of the insurance company, with the policy value serving only as collateral. McGill, Tr. 3783, 3785; Des-Rochers, Tr. 4860. The loans are not withdrawals of cash from the policy value itself. McGill, Tr. 3783, 3785; DesRoch-ers, Tr. 4860. 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. Id. In addition, there is no requirement that there be any pre-existing policy value for an insurance company to make a loan. Id. Even Kristie Sayre conceded a bank loan taken on the first day of the first year of a policy to pay the premium would be a real transaction. Kristie Sayre, Tr. 4196.

Moreover, the preponderance of the evidence indicates that first year, first day loans have been a common method of funding premiums for many years. Bossen testified it is common for a policyholder to obtain policy loans in each of the first three years of a policy in order to satisfy the payment of premiums. Bossen, Tr. 2128-29. Bossen characterized this type of insurance policy as minimum deposit insurance, which has existed since the late 1950’s. Bossen, Tr. 2129-30. Van Etten confirmed that minimum deposit insurance and first year, first day loans had been used since the 1970’s and early 1980’s, prior to development of the FA85 Policy. Van Etten, Tr. 2696. Also, Campisi testified that borrowing from the insurance company in order to pay premiums is a common feature of all life insurance policies. Campisi, Tr. 1008.

Moreover, the testimony of DesRochers, McGill, and Eisenberg reveals the commonality of first year, first day loans. Although he could not cite to a particular policy, DesRochers repeatedly testified first year, first day policy loans are common in the life insurance industry. Des-Rochers, Tr.1975-76, 4916, 4917. Further, although McGill, a government expert, testified he knew of no other policy that allowed first year, first day policy loans, McGill, Tr. 4851-54, he conceded it is common for first year policy loans to be granted on half the policy value if half of the premium is paid in cash. 27 McGill, Tr. 3817-21. In addition, he agreed that several leading insurance texts discuss the commonality of first year, first day loans in the life insurance industry. Id.; R610; R611; R644. Likewise, although Eisenberg testified that although he had never previously worked with first year, first day loans, he stated they are “fairly common.” Eisenberg, Tr. 2425-26.

Additionally, Camelot cites several cases in which a taxpayer took out first year, first day policy loans. See Shirar v. Comm’r, 916 F.2d 1414, 1416, 1418 (9th Cir.1990); Woodson-Tenent Lab., Inc. v. United States, 454 F.2d 637, 639 (6th Cir.1972); Campbell v. Cen-Tex, Inc., 377 F.2d 688, 689 (5th Cir.1967); Priester Mach. Co. v. United States, 296 F.Supp. 604, 605 (W.D.Tenn.1969); Golsen v. United States, 80 USTC (CCH) ¶9741, 1980 WL 4732 (Ct.Cl.1980) (“Golsen II”); see also Rev. Rui. 72-609 (1972). Although none these cases passed on the issue of whether first year, first day policy loans are shams in fact, they are instructive of the fact that first year, first day policy loans have been around the life insurance industry for many years.

In conclusion, Camelot has proven, by a preponderance of the evidence, that Camelot’s first year first day policy loans are within the norms of the life insurance industry.

(2) Risk Characteristics of Loans

The IRS asserts the policy loans lack the economic risk characteristics generally present in life insurance policy loans. Most commercial lending transactions have seven types of risk characteristics: (i) market, interest rate and reinvestment risk; (ii) liquidity or marketability risk; (iii) volatility, timing, and call risk; (iv) sector or event risk; (v) credit or default risk; (vi) maturity risk; and (vii) legal or operational risk. James W. Hoag (“Hoag”), Tr. 4564; G1081. The IRS risk characteristic position is weakened because the evidence shows it is standard for life insurance policy loans, including the Camelot COLI VIII policy loans, to lack all of these types of risk.

(a) Market, Interest Rate, and Reinvestment Risk

Market, interest rate, and reinvestment risk are the risk that market interest rates will vary from the interest rate on a loan. Hoag, Tr. 4565; G1081. From a lender’s perspective, this type of risk describes the risk that prevailing market rates for investments will exceed the interest rate on a borrower’s loan. Hoag, Tr. 4565. From a borrower’s perspective, this type of risk describes the risk that prevailing market rates for investment will drop below the interest rate on the borrower’s loan. Hoag, Tr. 4565. In a life insurance policy loan context, this type of risk is also referred to as disintermediation risk, or the risk that policy owner will be able to borrow funds at a rate which is less than the rate the insurance company could invest the funds. Puglisi, Tr. 4343-46. This risk is present in many life insurance policies that have fixed loan interest rates. Hoag, Tr. 4568-69.

As the IRS contends, the evidence demonstrates market, interest rate, and reinvestment risk is absent from the Camelot COLI policy loans because of the fixed 1% spread between the loan interest rate and the loaned crediting rate. As the IRS’s expert financial economist, Hoag, explained:

[T]he fixed spread at the core of the program, between the crediting and loan rate, in this case fixed at 1 percent, eliminates market, interest and reinvestment risk. You do not have to worry about the interest rate fluctuating because you always will be paying a fixed 1-percent spread for the privilege, if it is a privilege, of having the policy loan. There’s no reinvestment risk, because if interest rates change, in fact, you still get to pay your 1 percent cost in doing this. So none of the market-related risks exist.

Hoag, Tr. 4571; see also R238/G740 at 16— 17. Hoag’s assessment of this type of risk is supported by significant other testimony and evidence. Eisenberg, Tr. 2299-01, 2309-11; Van Etten, Tr. 2841-44; Puglisi, Tr. 4350-53, R188/G739 at 10-11; G55, Bates NJ649; G693, Bates 6170; G852, Bates 027016. Indeed, neither party has pointed to and the Court has not located any evidence to the contrary. 28

Moreover, the evidence also establishes that fixed spreads, which eliminate this category of risk, have been common features of life insurance policy loans for many years. DesRochers, Tr. 4869-70, 4872-73; Van Etten, Tr. 2794-95, 2795-96; Rogers, Tr. 129, 133-34, 144; Campisi, Tr. 1106-07; R15 at Bates CAM 004507; R557. Additionally, the IRS in a Technical Advice Memorandum has recognized that fixed spreads are “common with policy loans.” Tech. Adv. Mem. 98-12-005, at 48 (1998). Finally, in several reported cases, the taxpayer took out life insurance policy loans with a fixed spread. See Woodson-Tenent Lab., Inc., 454 F.2d at 639; Goldman v. United States, 403 F.2d 776, 778 (10th Cir.1968); Cen-Tex, Inc., 377 F.2d at 689; Priester Mach. Co., 296 F.Supp. at 605; Golsen II, 80 USTC (CCH) ¶ 9741, 1980 WL 4732. Although none of these cases addressed the issue of whether the fixed spread made the loans factual shams, they are instructive as to the existence of fixed spreads in the life insurance industry. Because the evidence overwhelmingly demonstrates other policy loans had fixed spreads, Camelot’s policy loans have the same lack of market, interest rate, and reinvestment risk characteristic of many other life insurance policy loans and is therefore within life insurance industry norms.

(b) Marketability or Liquidity Risk

Marketability or liquidity risk is the risk of the lender being unable to sell the loans to another party for their value. Hoag, Tr. 4565-66. Hoag testified that, with Camelot’s policy loans:

There’s no issue about liquidity or marketability because, in plain point of fact, you have a constant 1-percent spread and there’s no marketing going on, no fluctuation in the value to require an exchange.

Hoag, Tr. 4571. The Court has found no evidence in the record to contradict Hoag’s testimony. Because there is substantial evidence it is common for life insurance policies to use a fixed spread to eliminate this type of risk, the lack of liquidity or marketability risk of the Camelot policy loans is not unusual.

(c) Volatility, Timing, or Call Risk

Volatility, timing, or call risks are closely related to market and interest rate risk “because they depend on the fluctuations in interest rates rather than the level of interest rates.” Hoag, Tr. 4569. According to Hoag, the Camelot policy loans have “[n]o volatility, because the spread is, again, fixed and guaranteed.” Hoag, Tr. 4571. It follows that because fixed spreads are common, the lack of this type of risk with respect to the Camelot policy loans does not deviate from the norms in the industry.

(d)Sector Risk

Sector risk is the risk that one type of investment’s interest rates will fluctuate differently from another type’s. Hoag, Tr. 4566. Hoag testified Camelot’s policy loans had “no sector event risk” because of the fixed 1% spread. Hoag, Tr. 4571. Again, because fixed spreads are acceptable, the lack of sector risk is normal for a life insurance policy loan.

(e)Credit or Default Risk

Credit or default risk is the risk a borrower will default on a loan or be a credit risk. Hoag, Tr. 4566^1567. The evidence is undisputed that virtually all life insurance policy loans, including Camelot’s policy loans, have no credit or default risk because the loans are fully secured by virtue of the policy value serving as collateral. Hoag, Tr. 4568, 4574-76; R238/G740 at 17; Eisenberg, Tr. 2301-02; 2317-20; Puglisi, Tr. 4346^8, R188/G739 at 8-9; G852 at Bates 027016. Necessarily, the Camelot policy loans have the same lack of credit or default risk as virtually all types of policy loans.

(f)Maturity Risk

Maturity risk is the risk of when the loan will mature and what it will be worth at that time. Hoag, Tr. 4567. Because almost all policy loans, including Camelot’s policy loans, are linked to the death of the insured, there is no maturity risk. Hoag, Tr. 4568.

(g)Legal or Operational Risk

Legal or operational risk is a risk of: the law changing in between or new taxes being imposed, and on the operational side, whether or not certain payments happen as they’re proposed in the — in the loan project.

Hoag., Tr. 4567. Typical life insurance policy loans, including Camelot’s policy loans, have no operational risk because there is no third party involved to create this type of risk. Hoag, Tr. 4568. However, Camelot’s policy loans, like any tax advantaged transaction, have a legal risk of a tax law change. Van Etten, Tr. 2842; G852 at Bates 027016; G693 at Bates 6170.

In conclusion, Camelot’s policy loans have the same risk characteristics as typi cal life insurance policy loans. Because the risk characteristics do not vary from norms in the life insurance industry, the risk characteristics do not indicate the policy loans are factual shams.

(3) Artificially High Interest Rates

The IRS makes three arguments as to why Camelot’s policy loan interest rates indicate the policy loans were factual shams. First, the IRS asserts the policy loans are not real because Camelot elected the highest possible interest rate in a collusive, non-arms length transaction. In ordinary commercial lending transactions, interest is a cost borne by the borrower in exchange for the use of funds. Hoag, Tr. 4578. Ordinarily, if given the choice, a borrower will seek to minimize and the lender will seek to maximize the interest paid on a loan. Hoag, Tr. 4578. Although life insurance policies may give policyholders a choice between a fixed and variable loan interest rate, McGill, Tr. 3738, there is no evidence that policyholders ordinarily elect the highest possible interest rate. In contrast to ordinary lending practice and life insurance practice, Camelot was given a choice of a fixed rate and two variable rates and always chose the highest possible rate. Rogers, Tr. 291-293, 393-394, 396; Campisi, Tr. 1349-52; Hoag, Tr. 4577-81, 4589-91, R238/G740 at 17. Thus, the evidence demonstrates Camelot, by choosing the highest possible loaned crediting rate, did elect the highest possible loan interest rate in a collusive, non-arms length transaction.

Second, the IRS asserts the inflated loan interest rates selected by Camelot were excessive when compared to the essentially riskless nature of the policy loans. In general, the more risky an investment, the higher the interest rate. Moody’s rates investment grade corporate debt instruments as Aaa, Aa, A, or Baa, with Baa being the riskiest and having the highest interest rate. G710 at 4. “Bonds which are rated Aaa are judged to be highest quality. They carry the smallest degree of investment risk.” G710 at 4. On the other hand:

Bonds which are rated Baa are considered as medium-grade obligations (i.e., they are neither highly protected nor poorly secured). Interest payments and principal security appear adequate for the present but certain protective elements may be lacking or may be characteristically unreliable over any great length of time. Such bonds lack outstanding investment characteristics and in fact have speculative characteristics as well.

G710 at 4. The Moody’s Corporate Average is the composite average of these investment grade interest rates. G710 at 4; McGill, Tr. 3719-20, R163/G737 at 23; Pu-glisi, Tr. 4336-38, R188/G739 at 6-7.

The government witnesses differed as to an appropriate variable loan interest rate for a riskless policy loan such as the one in this case. Puglisi opined Camelot’s policy loans should have a loan interest rate approximating Moody’s Aaa rate. Puglisi, Tr. 4348, R188/G739 at 8-9. Bartlett and McGill testified an appropriate interest rate for Camelot’s policy loans would be the Moody’s Corporate Average. McGill, Tr. 2468; Bartlett, Tr. 4444. In addition, there is some evidence that other life insurance policies have used Moody’s Baa as the variable policy loan interest rate. Campisi, Tr. 1139; Van Etten, Tr. 2707.

By electing the highest possible Baa enhanced loaned crediting rate, Camelot automatically selected the highest possible loan interest rate. The following table illustrates how Camelot’s policy loan interest rate significantly exceeded the “appropriate” Moody’s Corporate Average and Moody’s Baa rates over the first seven policy years: 29

Policy Moody’s Moody’s Camelot’s Loan Year Corp. Avg. Baa Interest Rate

1 9,31% 9.81%12.99%

2 10.11% 10.74% 14.15%

3 8.98% 9.49%12.59%

4 8.38% 8.84% 11.79%

5 6.95% 7.31% 9.97%

6 8.77% 9.20% 12.23%

7 7.32% 7.75%10.49%

Stip. ¶¶ 167, 169, 171, 173-76; Moody’s Industrial Manual, v.l at a7S (1999).

There was no evidence that Camelot’s Baa enhanced loan interest rate, which regularly exceeded even Moody’s Baa by 2.74% to 3.41%, was acceptable or appropriate. Rather, this interest rate was a rate appropriate for high risk non-investment grade junk bonds, not to relatively riskless policy loans. Puglisi, Tr. 4339-42. The Court has no trouble concluding the evidence shows Camelot’s elected variable policy loan interest rate was excessive in light of the risk characteristics of the policy loans.

In conclusion, even though Camelot’s policy loan interest rate was set at the highest value possible in a collusive transaction and was excessive compared to the riskless nature of the policy loans, this is not enough for the Court to find the policy loans were factual shams. 30 Despite the inflated interest rate, Camelot did actually pay that inflated interest rate on the policy loans. Although the choice of the highest possible interest rate may be problematic from an economic substance point of view, it does not, by itself, make the policy loans shams in fact.

(4) Backdating of Loans

The February 16, 1990 Prepayment Agreement was an agreement to purchase the COLI VIII policy. The consideration was a relatively small amount of premium to secure the policy while Camelot and MBL worked out the details through the underwriting process. Rogers, Tr. 236; Campisi, Tr. 1196-97. Once finally purchased, Camelot’s COLI policy was considered to be in force as of the date of the Prepayment Agreement. Campisi, Tr. 1199-1200. On April 30, 1990, Camelot finalized the terms of its COLI life insurance policy, including the census of covered employees and the premium and policy loan amounts. Stip. ¶¶ 61, 64; Rogers, Tr. 268-82.

The government asserts the first year policy loan was not real because it was backdated and deemed to have been made on February 16, 1990, prior to the date it was actually made. There is no evidence that this practice violates accepted practices in the life insurance industry. To the contrary, a prepayment agreement generally defines the terms of the life insurance contract from the time of initial payment until delivery and acceptance of the contract. J27. The insurance company begins the underwriting process, which defines the terms of the policy, after the prepayment agreement is executed. Campisi, Tr. 1195-1200; Van Etten, Tr. 2803-04, 3007-09, 3226-28. When underwriting is complete, the policyholder makes all payments, including those made by policy loans, as of the date of the prepayment agreement. Campisi, Tr. 1195-1200; Van Etten, Tr. 2803-04, 3007-09, 3226-28. Because it is common for policy loans to run from the date of a prepayment agreement, Van Et-ten, Tr. 2804-05, Camelot’s first year, first day policy loan was not a factual sham by virtue of being treated as made on the date of the Prepayment Agreement.

(5) MBL’s Failure to Restrict Policy Loans During Rehabilitation

The IRS asserts the loans are not real because, when MBL went into rehabilitation due to financial difficulties, a court order put restraints on MBL making policy loans on all life insurance policies except those based on form FA85, including COLI VIII. Van Etten testified MBL did not treat the COLI VIII policies any differently from other policies, in that it allowed all premiums to be settled by policy loans if they could have been settled by loans in the past. Van Etten, Tr. 3112-14. However, Van Etten conceded his testimony was inconsistent with a state court order and internal MBL documents which indicated MBL could make policy loans only on the policies based on form FA85. Van Etten, Tr. 3109-14; G494 at Bates 003963; G495 at Bates 002029; G503 at Bates 003994. Thus, the preponderance of the evidence demonstrates MBL was permitted to treat form FA85 policy loans differently from other policy loans during rehabilitation. Nonetheless, the Court, on the record, is totally uninformed as to why the FA85 policy loans were treated differently and declines to speculate as to the reason. That being the case, the differential treatment is insufficient for the Court to find the policy loans to be shams in fact.

(6) Only 40% of Interest Paid in Cash

During the first eight years of Camelot’s COLI policy, Camelot incurred approximately $31.3 million in policy loan interest and paid about $12.5 million in cash via wire transfers to MBL and Hartford. Sayre, Tr. 4077-79; G973. In policy years one through three, the remainder of the accrued interest was paid via additional policy loans. G982. In policy years four through seven, the remainder of the interest was paid via partial withdrawals. 31 G982. The IRS asserts the policy loans were not real because only 40% of the accrued interest was paid in cash. However, absent any evidence that accrued interest must be paid only in cash, this fact alone does not make the policy loans factual shams.

Considering the characteristics of the policy loans in combination, the Court concludes Camelot’s COLI VIII policy loans were not factual shams. The evidence shows that it is acceptable practice in the life insurance industry: to make policy loans on the first day of each policy year; for policy loans to be essentially riskless; and to backdate the loans to the date of the prepayment agreement. There was also no evidence that it is unacceptable to pay only 40% of accrued interest in cash. The excessively high loan interest rates and, to a far lesser extent, the differential treatment of the COLI VIII policy loans during MBL’s rehabilitation provide some sparse indication the policy loans were not real. However, considering all of the factors together with the fact that Camelot actually had its death benefit payments reduced by the policy loans and accrued interest, the preponderance of the evidence persuades the Court that Camelot’s COLI VIII policy loans were real transactions, not factual shams.

b. Loading Dividends

The IRS challenges the loading dividends as being factual shams. Unlike the transaction involving the policy loans in years one through three, the netting transactions involving the loading dividends in years four through seven were cleverly designed accounting entries which occurred on paper but not in reality. A close examination of what occurred reveals that the loading dividends were not dividends, but rather paper transactions that had the sole purpose of reducing the amount of annual premiums that Camelot has to pay in cash. Bartlett, Tr. 3400-01, G736 at 5-6,19-20.

At first blush, this circular netting transaction involving paper loading charges, paper dividends, and paper premiums may appear to be indistinguishable from the seemingly similar circular netting transaction involving policy loans and premiums. However, they are different. In the netting transaction involving the policy loans and premiums, the policy loans are made by the insurance company, while the premiums originate from the coffers of the policyholder, in two distinct transactions which occur simultaneously. 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 sourced from another component in a circular transaction, the entire transaction is simply a paper ruse. 32

Eisenberg, the designer of the policy, acknowledged the annual premium, the expense charge, and the dividend, are created in a circular transaction with each one providing the basis for generating the other:

Q. You have. So what we have — correct me if I am wrong — we have the source of the dividend being the guaranteed expense charge, the source of the guaranteed expense charge is the premium, the source of the premium, in large part, is the dividend; correct?

A. I’ve got to think through this. The premium gets paid, that’s true. A 95-percent expense charge comes out, so 5 percent of the premium goes into the cash value calculation. 95 percent of the premium goes into the profits of the carrier, the surplus. 33 The actual expenses are, let’s say, 5 percent of premium. We now have 90 percent of premium as a profit to the insurance company and the surplus is taken out immediately, credited back as a dividend to the policyholder.

Eisenberg, Tr. 2505.

Bartlett, an IRS actuarial expert, explained the mechanics of this paper circular transaction in much the same way:

Well, the series of steps I’ve already alluded to where, in effect, the premium was designed in anticipation of this dividend and the premium would be paid by the policyholder, which would instantaneously create the, nominally speaking, the funds that would be remitted back to the policyholder in the form of the so-called loading dividend.

Bartlett, Tr. 3399. Camelot exhibits R573 and R574 similarly illustrate how the loading dividends arise from expense loading charges in a circular transaction.

In his expert report and testimony, Bartlett adopted the deposition testimony of Stan Brallier, a product analyst for a COLI brokerage firm, who aptly described this circular transaction as a “charade” designed to reduce the amount of premium that must be paid in cash:

From a product pricing perspective, MBL assumes that 95% of the annual premium will be required for policy expenses. This generates an inflated premium per thousand of insurance. In actuality, expenses are on the order of 5% of premium. Therefore, when the “expected” expenses do not materialize, MBL returns the excess charge to the policyholder through an inflated dividend in years four through seven. What this charade accomplishes is to return the majority of the premium to the policy holder, reducing the net cash outlay without loans.

G736 at 19-20 (emphasis added); see also Bartlett, Tr. 3400-01. Indeed, Endorsement EFA-4a refers to the loading dividends used to offset payment of premiums as dividends paid under a “Premium Reduction Plan,” which states “[the insurance company] will use dividends toward the payment of premiums.” J189 at Bates 001090.

Further, in contrast to the policy loans, the loading charges and loading dividends function in a manner which departs from many, if not all, accepted practices in the life insurance industry because: (1) the magnitude of the loading charges was excessive; (2) the loading dividends were not paid from surplus; (3) the loading dividends were paid at the beginning, rather than the end, of the policy year; (4) the loading dividends were not sourced from all experience factors; (5) the loading dividends were virtually guaranteed; and, (6) MBL and Hartford failed to properly account for the loading dividends. These characteristics will be discussed in turn.

(1) Magnitude of Loading Charges

The IRS asserts the magnitude of the loading charge in years four through seven is grossly excessive when compared to the magnitude of the anticipated expenses and has the sole purpose of providing the phantom funds for the phantom dividend. Camelot counters that it is acceptable life insurance practice to arbitrarily set loading charges at any level, including one high enough to provide for payment of dividends. The Court concludes the preponderance of the evidence supports the magnitude of the loading charges as being grossly excessive when compared to the magnitude of the anticipated expenses.

It is commonly accepted for life insurance policies to have loading charges to cover the insurance company’s expenses in administering the policy. Bartlett, Tr. 3373, G736 at 13; McGill, Tr. 3621-24, R163/G737 at 30 & Appendix 2. It is also well accepted to make the magnitude of the loading charge greater than the anticipated expenses in order to provide a reasonable margin of safety to cover unpredictable expenses, a permanent contribution to surplus, and a minimum cushion for dividends in the event of favorable experience. Bartlett, Tr. 3383-88, 3392, G736 at 13; McGill, Tr. 3621-24, R163/ G737 at 30 & Appendix 2. However, the magnitude of this cushion in the loading charge should reasonably relate to the risk of the insurance company incurring higher than expected expenses. Bartlett, Tr. 3392-95, G736 at 16; McGill, Tr. 3624-29, R163/G737 at 34-35; Eisenberg, Tr. 2246.

Camelot points to McGill’s concessions, on cross examination, that there is no rule or yardstick to determine what is a reasonable relationship between the magnitude of the loading charge and anticipated expense and that insurance company management has the discretion to arbitrarily determine the magnitude of loading. McGill, Tr. 3852-53. In addition, McGill concurred with a statement from MCGILL’S LIFE INSURANCE that:

The extent to which the loading formula will be used to create dividends will depend upon broad managerial viewpoints. It is primarily a question of whether the company wants to be a high-premium, large-dividend company or a low-premium, small-dividend company.

McGill, Tr. 3625; R140 at 257. 34

However, these statements are taken out of context. McGill more fully explained, as was discussed in MCGILL’S LIFE INSURANCE, arbitrary loading margins are only used when expenses are ill-defined and management wishes to charge a large loading charge and return the unused portion as a dividend at a later time:

Q: [quoting MCGILL’S LIFE INSURANCE] “Theoretically, at least, the loading charge in the dividend formula should be based on the same type of cost analysis that should have preceded the computation of the gross premium. If expenses are analyzed carefully and accurately in the preparation of gross premium scales and are apportioned equitably over the various plans, ages of issues and durations [sic?] through the loading formula, there is no reason why the same expense charges, suitably modified to reflect subsequent experience, could not and should not be used in the dividend formula. Some companies, however, arrive at the loading in a rather arbitrary fashion and justify their actions on the ground that any inequities in the premium loading can be corrected through the dividend formula. This merely postpones the day of reckoning, [since] the expense charge of the dividend formula, if it is to correct the inequities in the loading formula, must be calculated with all the attention to cost-accounting principles and the exercise of judgment that would have had to go into the calculation of a theoretically correct gross premium. There may be an advantage in the latter procedure, however, in that the cost calculations are made on the basis of actual, rather than projected, experience.”

What did you mean in that paragraph?

A. Well, I was, I think all actuaries— and I am not an actuary — but all actuaries recognize that the most difficult part of the pricing process is arriving at a— an allowance for loading — for expenses. I’m sorry. For expenses. There are all sorts of problems and cost allocations. And while the companies are under obligation to be — to deal fairly with all classes of policyholders, and they do try to allocate expenses in an equitable manner, a mutual company can have a little more freedom in determining the loading for expenses because they can, at the end of the day, at the end of the year, through the dividend formula, they have that mechanism to make adjustments in the cost to the dividend formula. And that was all I was saying here, that what might look like it was arbitrary in the beginning is trued up at the end of the year based on experience, through dividends.

McGill, Tr. 3627-28. Moreover, he explained that, even when insurance companies build in margins to their loading dividends, a generous margin would be “5 to 10 percent perhaps of the gross premium.” McGill, Tr. 3626.

Camelot also points to Bartlett’s testimony that he knows an excessive loading charge when he sees it and that a loading charge of more than 30% to 40% of the annual premium would be excessive, Bartlett, Tr. 3585-88. Camelot also directs the Court’s attention to Bartlett’s admission that his opinion differs from the opinions of Messrs. Berkowitz, Sunder-geld, and Odell, Camelot’s expert witnesses, and the expert witnesses for American Electric Power in a closely related COLI tax case. Bartlett, Tr. 3589. However, since none of these witnesses testified at this trial to contradict Bartlett’s testimony, and because the Court finds Bartlett credible, the Court credits his opinion on this matter.

In this case, the magnitude of the loading charges in the fourth through seventh policy years is extreme when compared to the anticipated expenses and risks. For these years, the anticipated expenses are around 7.90% of the premium, while the loading charges are 95% of the premium, yielding a margin (the difference between the loading charge and the anticipated expense) of about 87.70% of the premium. G877. The anticipated expenses, comprising compensation paid to IAS and IAD, commissions paid to Newport, state premium taxes, and federal taxes paid by the insurance company, were fixed by contract or statute. Van Etten, Tr. 2724-25; G136 at Bates 009540-009541; Bartlett, G736 at 13-14. It follows there was minimal risk that actual policy expenses would be greater than anticipated. Bartlett, Tr. 3375-80, G736 at 13-14; McGill, Tr. 3632-36, R163/ G737 at 34-35. Indeed, the actual expenses never exceeded 10% of the premium. Bartlett, Tr. 3378; McGill, Tr. 3632-36, R163/G737 at 34-35; G877 (actuarial memorandum). The conclusion is inescapable that the premium loading was substantially higher than anticipated expenses in the fourth through seventh policy years. G136 at Bates 009540; McGill, R163/G737 at 34-35; Eisenberg, Tr. 2480-83; Van Etten, Tr. 2777-78.

Given the low risk of greater than anticipated expenses, the magnitude of the loading charges was “grossly excessive.” Bartlett, Tr. 3359; see also id. 3373-74, 3388, G736 at 15-18; McGill, Tr. 3617-18, 3638-3640, R163/G737 at 37. The 87% margins were substantially greater than the 5-10% margins McGill called generous or the 30-40% margins Bartlett called excessive. 35 McGill, Tr. 3626; Bartlett, Tr. 3585-88. Instead of providing a margin for risk, these excessive loading charges were built into the policy only to fund the “extreme dividend patterns” as refunds of premium expense charges. G75; see also G89 at Bates H0053; G102 at Bates ICMG 289; Van Etten, Tr. 2901. As Bartlett testified:

[T]he purpose of setting the loading at the level was not to perform the functions ... loading traditionally was expected to fulfill but was, rather, intended to provide a source of funds to, in effect, minimize the cash [outlay] required on this policy in years four through seven.

Bartlett, Tr. 3391.

Indeed, Eisenberg conceded the loading charges were never intended to be close to actual expenses. Eisenberg, Tr. 2242^44. Rather, he stated the loading expenses were used as a source for dividends to minimize the cash outlay for premiums in years four through seven. Eisenberg, Tr. 2483; 2488; 16; see also Van Etten, Tr. 2777-78, 2898-2900. Eisenberg’s testimony illuminates the excessive nature of the loading charges:

So we never intended — I mean, obviously, we never expected to have true expense of 95 percent. We’re not pretending that we tried to do that.... [Premiums ... come from different sources, cash being one of the sources. There was a little bit of cash being paid and, you’re right, it wasn’t a lot. A little bit being paid in years four, five, six and seven. The other source of funds for consideration of the premium was going to be dividends, and you can see from here that we expected to have some very large dividends.... So that is definitely a design consideration, was to have a high dividend to provide funds to help pay premiums.

Eisenberg, Tr. 2242-44.

Finally, Eisenberg and Van Etten both conceded they had never encountered another life insurance policy with this magnitude of loading. Eisenberg, Tr. 2244^45; Van Etten, Tr. 2901. Given this evidence, the Court concludes the magnitude of the loading dividends excessively departed from commercial norms, and is indicative of the dividends being factual shams. 36

(2) Dividends Not Paid From Surplus

In traditional dividend practice, dividends to policyholders are a distribution of the insurance company’s prior accumulated divisible surplus. Van Etten, Tr. 2906-07. Near the end of each calendar year, insurance company officers determine how much surplus money the company has. Id.; Bartlett, Tr. 3397. The amount of surplus funds is based on past experience over that year plus a prediction about experience for the remainder of the year. McGill, Tr. 3640, 3654-55, R163/G737 at 39, 49-50. The Board of Directors determines what portion of the total surplus should be distributed as dividends (“divisible surplus”) and what portion should be retained for general corporate purposes. Van Etten, Tr. 2907; Bartlett, Tr. 3397; McGill, Tr. 3653-54.

In contrast to accepted dividend practice, Camelot’s loading dividends were not sourced from MBL’s surplus. Van Etten, Tr. 2911-13, 2932; G89 at Bates H0040. Rather, the source of the loading dividend was the anticipated excessive loading charge which, upon deemed payment of the premium, would be instantaneously returned as an offset against payment of the premium. Eisenberg, Tr. 2242-43, 2276-78, 2483; 2488; 2504-06; Van Etten, Tr. 2777-78, 2898-2900, 2922-30, 2930-32; Bartlett, Tr. 3413-14; McGill, R163/G737 at 49-50, 54; G516 at Bates MBL 3211-3213. Viewed in this manner, the so called “dividend” could be more accurately designated as a rebate or reduction of premium or credit to premium.

Eisenberg and Van Etten would have the Court believe that surplus could somehow be instantaneously created by payment of premium and removal of the 95% expense charge, while at the same time be instantaneously returned to Camelot as a dividend. Eisenberg, Tr. 2234; Van Et-ten, Tr. 2769-70, 2930-31, 2934-35, 3191, 3205-09, 3243-47. This instantaneously created surplus runs counter to the norms in the life insurance industry. With all due deference to Messrs. Eisenberg and Van Etten, the type of surplus envisioned by them exists only in their creative imaginations. The preponderance of the evidence supports the conclusion that Camelot’s loading dividends were not paid from the surplus of the insurance company.

Camelot, recognizing the weakness of its position, also contends the loading dividends need not be sourced from divisible surplus. Camelot refers to Actuarial Standard of Practice (“ASOP”) No. 29, which defines “policyholder dividend” in the alternative as “non-guaranteed returns of premium or distributions of surplus.” R598 at 23 (emphasis added); see also R601 at 21. Van Etten, McGill, and Bartlett concurred with this definition. Van Etten, Tr. 3145-46; McGill, Tr. 3828; Bartlett, Tr. 3484. In addition, McGill concluded there is no legal requirement that dividends be sourced from surplus. McGill, Tr. 3847-48. Bartlett likewise admitted it is common for insurance companies to pay dividends to policyholders even if they do not contribute to surplus. Bartlett, Tr. 3532. However, this argument fails because, as discussed infra, section III.A.2.b.(5), the loading dividends are “virtually guaranteed,” so that they cannot be “non-guaranteed” returns of premium, as provided in ASOP No. 29. Therefore, the preponderance of the evidence supports the conclusion that the loading divi dends were neither paid out of surplus, nor a non-guaranteed return of premium.

Dividends are generally distributed retrospectively at the end of a policy year. McGill, Tr. 3655-56, R163/G737 at 52; Bartlett, Tr. 3411-12; Eisenberg, Tr. 2231-33. McGill explained the reason for distributing dividends at the end of the policy year is so that they reflect, as closely as possible, the insurance company’s actual experience over the year:

Well, that’s — it means that you don’t— you don’t distribute future surplus. You distribute only what has been accumulated in the past. It doesn’t mean that in any particular year, the surplus that you distributed was precisely what the operating gains were for that particular year. There could be more or less or any number of reasons why that would be true. But the basic thing is that it always represents — I mean, dividends always represent a distribution of moneys that have accumulated through the past experience of the company, not future experience.

McGill, Tr. 3654. Camelot points to Bartlett’s testimony that no actuarial standard prohibits payment of dividends at the beginning of a policy year. Bartlett, Tr. 3532. However, paying dividends at the beginning of a year violates the contribution principle, which requires dividends to be apportioned among the policyholders in proportion to each one’s prior contribution to the divisible surplus. Bartlett, Tr. 3411-12; McGill, Tr. 3643; Eisenberg, Tr. 2232-33. Accordingly, the preponderance of the evidence establishes that dividends are not distributed prospectively on the first day of a policy year.

Nonetheless, the loading dividends were paid prospectively on the first day of each policy year, violating the accepted practice of only paying dividends retrospectively at the end of a policy year. Stip. ¶ 132; Ei-senberg, Tr. 2231-35, 2489-90; Van Etten, Tr. 2863-64; Bartlett, Tr. 3412-13, G736 at 23; McGill, Tr. 3655-56, 3675-78, 3684, R163/G737 at 51-52, 54-55; G89 at Bates H0053. Since all of the expenses were contractual, that portion of the loading dividends to be used for expenses was based on anticipated experience, which allowed payment of a limited prospective dividend at the beginning of the policy year. Eisen-berg, Tr. 2231-35; McGill, Tr. 3675-76. Indeed, a dividend was paid to Camelot, and credited to policy value, on the first day of the first year of the policy, having been declared by MBL the prior November, before Camelot even purchased the COLI policy! 37 Van Etten, Tr. 2905-06; Stip. ¶¶ 119-20; Eisenberg, Tr. 2490-94. Eisenberg, the principal designer of the policy, found this characteristic of the dividends to be “innovative.” Eisenberg, Tr. 2672. It follows, the preponderance of the evidence demonstrates the loading dividend was prospectively paid on the first day of the policy year, contradicting the accepted life insurance practice of paying dividends retrospectively based on prior experience.

(4) Sources of Dividends

Normally dividends are paid out of divisible surplus based upon three experience factors: mortality rates, investment yields, and expense rates. McGill, Tr. 3643-44; Bartlett, Tr. 3410. Dividends are based on the aggregate surplus created by the combination of all of these factors, and cannot be based on only a single factor. Id. The expense savings from the loading charge is normally a small component of a dividend, when compared to mortality and investment experience. McGill, Tr. 3672-73. In violation of this accepted practice, Camelot was paid a loading dividend sourced from the loading charges at the beginning of the policy year and a separate dividend based upon mortality experience at the end of the policy year, while there was no dividend based on investment experience. McGill, Tr. 8670-72, R163/G737 at 62; Bartlett, 3410. The amount of the loading dividend far exceeded the amount of any mortality dividends. McGill, Tr. 3670-72; G982.

Camelot points to a concession by Bartlett that no statutory, regulatory, or actuarial standard requires a single dividend based on all of these factors. Bartlett, Tr. 3528. However, Bartlett then explained:

Every single text that the Society [of Actuaries] uses with which I am familiar, including those that I have appended to my report, show the formula for dividends is as [sic] incorporating all three elements.

Bartlett, Tr. 3528. Moreover, Bartlett had also stated a single aggregate dividend is implicitly required by ASOP No. 15, the contribution principle, which requires that dividends “be distributed among policies in the same proportion as the policies are considered to have contributed to divisible surplus.” R598 at 9; Bartlett, Tr. 3410-12. The reason for the practice of paying one aggregate dividend is so the insurance company, if necessary, has the freedom to use excess funds from loading charges to make up for bad mortality or investment experience. Bartlett, G736 at 22. Indeed, Bartlett stated he had never seen another policy that declares a dividend based on only one factor. Bartlett, Tr. 3412. The Court holds the preponderance of the evidence establishes the COLI VIII loading dividends violated the principle that dividends be based upon the aggregate of experience factors. Further the Court finds the norm in the life insurance industry is the loading charge provides only a small component of the dividend.

(5) Virtually Guaranteed

The IRS asserts that Camelot’s loading dividends were virtually guaranteed. It is an undisputed fact that payment of life insurance “dividends, by law, cannot be guaranteed by the carrier.” G161 at Bates M & R 008592; see also G85 at Bates AEP-MBL 02219; McGill, R163/G737 at 52; Bossen, Tr. 1774. In a predecessor to COLI VIII, Van Etten drafted a “Dividend Guarantee Endorsement,” which created expense charges equal to approximately 100% of the premium in policy years three and thereafter, and a guaranteed loading dividend equal to approximately 90% of the premium in policy years four through seven. G79; see also Van Etten, Tr. 2861-71; Eisenberg, Tr. 2463-70. 38 However, when it became clear to the designers that dividends could not be guaranteed under existing law and state regulations, McCamish insisted on finding a way to virtually guarantee payment of the dividend:

The second question centers around the “guarantee” of the dividend schedule. As we all know, the actual dollar dividend has been factored into the policy design ... Mutual Benefit has no financial risk whatsoever. Therefore, it is a totally regulatory issue. We simply must find a way to guarantee the dividend. Steve Eisenberg said that we can guarantee the roughly 80% of the dividend by the formula within the policy. It’s the other 20% that is causing problems. Here again, the premium amounts involved are such that a 20% unknown is totally unacceptable. This is particularly true in light of the fact that it is not a financial consideration on Mutual Benefit’s part. We simply must find a way to achieve the desired result.

G80 at Bates M & R 008334 (emphasis added, ellipsis in original); see also McCamish, Tr. 4806-07; 4808-09; Van Et-ten, Tr. 2886-87; G86 at Bates 003731.

The COLI VIII designers achieved McCamish’s goal of virtually guaranteeing the dividend by making three changes to the policy form. First, as previously discussed, the designers imposed extremely high expense charges in policy years four through seven. Eisenberg, Tr. 2477-78; Van Etten, Tr. 2922-26; G516 at Bates MBL 3208; G868 at Bates 1380. Second, they eliminated the “Dividend Guarantee Endorsement” but used the same formula in the dividend resolution submitted to the insurance company’s Board of Directors each year. Id. Finally, they made the loading dividend payable on the first day of each policy year, in part, so that, if the insurance company failed to pay the dividend, the policyholder theoretically would not have to pay the premium. G84; G130; G161; G317 at Bates 11202; Eisenberg, Tr. 2473-77; 2490-94, 2496-98; McGill, Tr. 3676-77; Bossen, Tr. 2102-05.

It is undisputed that, with these changes, the COLI VIII policy did not actually guarantee payment of a dividend. Rogers, Tr. 324; Bossen, Tr. 1755, 1774, 1778; Eisenberg, Tr. 2230, 2625; Van Et-ten, Tr. 2777; McGill, Tr. 3841. However, these changes had the effect of making the dividend “virtually guaranteed.” Bartlett, Tr. 3403-08, G736 at 21-22; McGill, Tr. 3657-58, 3676-77, 3679-82, R163/G737 at 55-57; Bossen, Tr.2098-2102, 2102-05; Ei-senberg, Tr. 2473-77; 2490-94, 2496-98, 2531-33; Van Etten, Tr. 2928-30; McCamish, Tr. 4810-12; G84; G130; G161; G162 at Bates 000981; G228; G296 at Bates 007262; G317 at Bates 11202; G367 at 3. As explained by Van Etten, “conceptually, in lieu of paying a benefit that was at one time labeled a guarantee[d] dividend, [premium expense] amounts in the same general range were incorporated in the formula for payment of dividends.” Van Etten, Tr. 2928-30. As explained by Eisenberg, “since the policyholder will know, prior to his premium payment, if Mutual Benefit is not maintaining the scale, the client need not pay the premium from which the dividend arises. Consequently, the client is completely protected from any reduction in dividends.” G161. In a memo commissioned by MBL, the law firm of Steptoe and Johnson found this policy design “in effect, guarantees substantial dividends.” G162 at Bates 000981.

MBL and IAS touted the virtual guarantee of the loading dividends in marketing the product to brokers and policyholders. G228, G161; G296 at Bates 007262; G317 at Bates 11202; G367 at 3; Bossen, Tr.2098-2102; Eisenberg, Tr. 2531-33; McGill, Tr. 3680-82; McCamish, Tr. 4810-12. For example, in a letter to the Newport Group, IAS stated:

As you know, by their nature, insurance dividends cannot be guaranteed. However, COLI V dividends are “virtually guaranteed” by an expense loading component. 39

G228 (emphasis added). Also, in an Eisen-berg letter forwarded by Bossen to a broker, Eisenberg explained how the product was designed to achieve the illustrated dividends:

As we have discussed on several occasions, the likelihood of dividends being paid on MBL’s COLI product are extremely high. The primary source of dividends is expenses. Since Mutual Benefit contracts out administration to IAS, MBL’s costs are contractual and cannot really fluctuate.... Since this dividend source is consistent with pricing assumptions as filed with the insurance department, this dividend source should not fluctuate. Since none of the illustrative dividend arises from a mortality source, the likelihood of the divi dends being paid as illustrated is quite high.

G367 at 3.

MBL even went so far as to promise one COLI policyholder that MBL would redesign the policies if it could not pay the projected dividends:

should there be any future uncontrollable circumstances that adversely affect the calculation of dividends, MBL will redesign or amend such affected policies ... to achieve substantially the same results....

G88 at Bates M & R 001854; see also Van Etten, Tr. 2886-95. 40

Camelot argues the loading dividends could not have been virtually guaranteed because the actual loading dividends were somewhat less than the projected loading dividends. Van Etten, Tr. 2775. Part of the reason for the difference is that, beginning in October, 1991, MBL deducted larger than anticipated expenses from the loading charge to account for the newly imposed federal DAC tax. 41 Eisenberg, Tr. 2624-25. However, to avoid MBL having to substantially deviate from the projected loading dividend schedule, MBL and Newport each absorbed about one third of the DAC tax (Newport’s absorption of its share of the DAC tax was accomplished by commission reduction), and passed the balance on to the policyholder. Campisi, Tr. 1260.

Indeed, despite the DAC tax, loading dividends over the first seven years of the Camelot’s policy closely matched the projections shown to Camelot prior to its purchase of the policy:

Projected Actual Loading Loading Difference Difference Year Dividend 42 Dividend (in Dollars) (Percent)

1 0 0 0 0.00%

2 0 0 0 0.00%

3 967 1,039 72 7.45%

4 12,641 12,506 (135) (1.07)%

6 12,557 12,487 (70) (0.56)%

6 12,472 12,354 (118) (0.95)%

7 12,458 12,131 (327) (2.62)%

Total 51,095 50,517 (578) (1.13)%

J44 at Bates 004937; J138 at Bates 006089 (000’s omitted). As shown in the above chart, over the first seven years of the policy, Camelot received only 1.13% less than the projected loading dividend. Despite the added DAC tax and the fact that MBL went through rehabilitation, MBL never missed a loading dividend payment on any of its 50 COLI policies. Eisenberg, Tr. 2676-77. This performance during a period of financial adversity, coupled with the incredibly close match between the projected and actual loading dividends, persuades the Court, by a preponderance of the evidence, that the loading dividends were “virtually guaranteed.”

(6) Accounting for Dividends

State accounting regulations require an insurance company to create a liability on its balance sheets for all dividends. Bartlett, Tr. 3415, G736 at 24; McGill, Tr. 3656-57, R163/G737 at 52. Neither MBL nor Hartford set up a liability on then-books for the loading dividends because they are a “refund of premium expense charges ... paid at the times premiums are collected.” G89 at Bates H0053; see also Van Etten, Tr. 2911, 2930, 2983-85, 2986-88; Bartlett, G736 at 24-25; McGill, R163/G737 at 55. 43 The failure to create a liability for the loading dividends, as required by state accounting rules, provides some indicia that the loading dividends were not real dividends. Bartlett, Tr. 3414-15, G786 at 24-25; McGill, Tr. 3684-85, R163/G737 at 55.

Considering these six characteristics together, the loading dividends lack indicia of real dividends because: (1) the magnitude of the loading charges was dramatically excessive compared to anticipated expenses; (2) the loading dividends were not paid out of surplus; (3) the loading dividends were paid at the beginning, rather than at the end, of each policy year; (4) the loading dividends were sourced from only prospective expense experience, and not from prior mortality and investment experience; (5) the loading dividends were virtually guaranteed; and, (6) no liability was set up for the loading dividends. 44

Despite this mountain of evidence, Camelot argues the loading dividends are real because they are “policyholder dividends” as defined in I.R.C. § 808, which defines “policyholder dividend” as “any dividend or similar distribution to policyholders in their capacity as such.” 26 U.S.C. § 808(a) 45 This argument fails because the § 808 definition of policyholder dividends rests on the assumption that the dividends are real transactions, and not factual shams. The definition of “policyholder dividend” as “dividend or similar distribution” implicitly assumes that the dividends would actually be real transactions. As discussed, supra, section III. A.2.(b), there is a mountain of evidence that Camelot’s loading dividends are factual shams. Camelot’s argument that § 808 can be used to determine whether the loading dividends are real is a disingenuous attempt at bootstrapping. Cf. United States v. Mitchell, 145 F.3d 572, 578-79 (3d Cir.1998). Essentially, Camelot argues that, if the loading dividend falls within the definition of § 808, it must be real, even though § 808 requires the policyholder dividends be real in the first place. This attempt at bootstrapping does not make the loading dividends any less a factual sham.

Given the overwhelming evidence, the Court holds the loading dividends in years four through seven were factual shams.

c. Partial Withdrawals

The IRS relies on Kristie Sayre’s testimony asserting the partial withdrawals are factual shams because, at the beginning of each policy year, there is zero net equity to support taking a partial withdrawal. Kristie Sayre, Tr. 4036-38, 4042-46, 4056, 4059-60, R314/G738 at 42-43, 47-48. The IRS contends the partial withdrawals, like the loading dividends, are illusory because they are both sourced from, and the basis for, the deemed payment of the premium and accrued loan interest. Kristie Sayre, Tr. 4036-38, 4042-46, 4056, 4059-60, R314/ G738 at 42-43, 47-48. The government’s contention fails because Camelot’s policy allowed Camelot to take “a partial withdrawal of the policy value,” and places no limit on the amount of policy value that may be taken as a partial withdrawal. J189 at Bates 001093 (Endorsement EFA 85-7); Bossen, Tr. 1743, 1758; DesRochers, Tr. 4859.

On the day before Camelot took each of the partial withdrawals in years four through seven, the policy had a net policy equity equal to zero. Eisenberg, Tr. 2225-26, 2518-22; Kristie Sayre, Tr. 3994-96, 4002-09, 4010-11, 4042, R314/G738 at 40; G792; G844 at Bates 5583; G848 at 2; G949 at Bates 015150; G972; G987. However, there was sufficient policy value, created by the payment of premiums in years one through three, plus the inside build-up on these premiums, to support taking the partial withdrawal. Moreover, each partial withdrawal of policy value was in an amount slightly less than the amount of outstanding accrued loan interest:

Year Policy Value Net Accrued Partial Partial Withdrawal as Percent _Equity Interest Withdrawal of Accrued Interest_

4 48,093,000_0 5.377,808 5,308,348_98.71%

5 47,684,000_0 5,032,552 4,986,947_99.09%

6 46,875,000_0 4,249,749 4,245,614_99.90%

7 47,737,000_0 5,209,413 5,203,549_99.89%

J93; J109; J115; J123; J128; G738 at 40. Because the policy had zero net equity, the partial withdrawals lowered the policy value and loan limit and were used to lower the loan balance by paying most of the accrued policy loan interest. Campisi, Tr. 1432; Eisenberg, Tr. 2495. The partial withdrawals did not go towards payment of the premiums in years four through seven. In addition, the partial withdrawals reduced the death benefits. Eisen-berg, Tr. 2495; DesRochers, Tr.1983. It follows, the partial withdrawals in years four through seven must also have been real. 46 DesRochers, Tr.1983.

The real partial withdrawals are different in kind from the sham loading dividends. The dividends were shams in fact, in large part, because they were created on paper in a circular netting transaction, where they were both sourced from, and the basis for, the premium payment. On the other hand, the partial withdrawals were sourced from the prior existing policy value but used to pay accrued loan interest. Because the partial withdrawals were sourced from the policy value, but applied to payment of loan interest, they were not created in a circular transaction, so as to make them factual shams.

d. Annual Premiums

In years one through three, the annual premiums were paid through a combination of about 88% policy loans and about 12% cash. Kristie Sayre, Tr. 3981-82; R314/G738 at 37; G970. The IRS contends the annual premiums in these years were not real because they were sourced from contrived policy loans that did not occur in reality. However, as discussed in detail, the policy loans in the first three years were real transactions. See, supra, section III.A.2.a. Since the policy loans were real, it stands to reason the annual premiums in years one through three were also real.

In years four through seven, the annual premiums were purportedly paid by a combination of about 12% cash and about 88% loading dividends. 47 Kristie Sayre, Tr. 3981-82; R314/G738 at 37; G970. The IRS asserts the annual premiums in these years are not real because they were created in a circular netting transaction by the sham loading dividends. As discussed, the loading dividends were not real. See, supra, section III.A.2.b. Since the loading dividends were the purported source of about 88% of the annual premiums and were purportedly sourced from the deemed payment of annual premiums, the purported payments of annual premiums in years four through seven were also factual shams.

The IRS contends none of the annual premiums were real because they were set at the highest possible level. It is undisputed Camelot’s policy was designed to and had a high premium per thousand, exceeding traditional whole life insurance policies by 250-300%. Eisenberg, Tr. 2200, 2432-33; Van Etten, Tr. 2719-20, 2740-42; Kristie Sayre, Tr. 3946-50, 3948, 3950, R314/G738 at 26-27. However, it is also undisputed the level of the premiums did not exceed the seven-pay test in I.R.C. § 7702A, which sets a cap on the level of life insurance premiums as compared to death benefits. 26 U.S.C. § 7702A; Stip. ¶¶ 93-95; Tr.1958; Eisenberg, Tr. 2200, 2202-03, 2214. Although compliance with § 7702A does not automatically make the premiums real, the fact that Congress authorized such high premiums is fatal to this IRS argument and is a strong indication the premium structure would be acceptable in the life insurance industry. Accordingly, the level of the premiums does not make them factual shams.

The IRS ignores the reality that life insurance premiums can, and often do, exceed the amount needed to fund the COI charges, actual expenses, insurance company profits, and death benefits. Indeed, Congress recognized this fact when it enacted I.R.C. §§ 7702, 7702A to place limits on the amount by which the premiums could exceed that needed to fund the death benefits. 26 U.S.C. §§ 7702, 7702A; Stip. ¶¶ 93-95; DesRochers, Tr.1958; Eisen-berg, Tr. 2200, 2202-03, 2214; DesRoch-ers, R340. It follows, even if the cash payments were sufficient to support the COI charges, actual expenses, insurance company profits, and death benefits, the fact that the annual premiums were far higher than required does not make them factual shams.

Camelot asserts the annual premiums were real because they comply with the caps in I.R.C. §§ 7702, 7702A. See Des-Rochers, R340. However, §§ 7702 and 7702A relate only to the magnitude of the annual premiums and policy value relative to the magnitude of the death benefits. 26 U.S.C. §§ 7702, 7702A. As discussed supra, section III.A.1., compliance with these code sections does not mean the premiums are automatically real, as that would eviscerate the sham transaction doctrine. See ACM Partnership, 157 F.3d at 245^46. Rather the court must evaluate whether the transaction is a sham apart from whether its form complies with provisions of the tax code. See id. at 246. As discussed, supra, section III.A.2.b., because the premiums in years four through seven were sourced from, and were the basis for, the loading dividends in a circular transaction, the premiums did not occur in reality, regardless of whether their purported magnitude complies with the requirements of I.R.C. §§ 7702, 7702A.

Finally, Camelot argues the annual premiums are real in all years because they were reported as income by MBL and Hartford on their financial statements and state and federal income tax forms. Having found the premiums to be real in years one through three, the Court need only address this argument as it relates to the premiums in years four through seven.

The evidence does show MBL and Hartford recorded the annual premiums each year on its books. Bossen, Tr. 2127-28. The evidence also indicates MBL and Hartford reported the annual premiums for purposes of state premium tax and the federal DAC tax. Bossen, Tr. 2130; Van Etten, Tr. 2724-26. However, in policy years four through seven, MBL and Hartford were also able to deduct the loading dividends from the premiums for purposes of calculating their state premium tax. Bossen, Tr. 2130. Since the loading dividends were offset against the annual premiums for tax reporting purposes in years four through seven, the reporting of premiums in those years does not alter the conclusion that the premiums were not real in those years. Similarly, the DAC tax would have been passed through in its entirety to Camelot. However, in order to preserve the accuracy of the projections furnished Camelot, MBL and Newport each absorbed one-third of the DAC tax with the remaining one-third passed through to Camelot. MBL’s decision to voluntary pay one-third of the DAC tax without recouping it from Camelot does not make the premiums real in years four through seven.

In conclusion, the preponderance of the evidence shows the annual premiums were real in years one through three, but were factual shams in years four through seven.

e. Considering All Features Together

Ultimately, the Court must consider all components of the Camelot COLI VIII plan in combination to determine whether the interest paid was real or a sham in fact. See ACM Partnership, 157 F.3d at 247. In years one through three Camelot paid real annual premiums, took out real policy loans which accrued real interest, and paid the real interest by a combination of cash and real policy loans. In years four through seven, Camelot paid about 99% of the real interest on the real policy loans with the real partial withdrawals and paid the additional approximately 1% by cash. But also, in years four through seven, the annual premiums were shams in fact because they were primarily funded by the sham loading dividends. Since arguably all of the accrued interest was paid through real transactions, Camelot’s interest deductions for the first three policy years were not the product of factual shams.

3. Sham in Substance/Economic Substance Doctrine

In ACM Partnership, the Third Circuit Court of Appeals held “pursuant to Gregory [v. Helvering, 293 U.S. 465, 55 S.Ct. 266], we must ‘look beyond the form of [the] transaction’ to determine whether it has the ‘economic substance that [its] form represents.’ ” 157 F.3d at 247 (citation omitted) (brackets in original). “[R]egardless of its form, a transac tion that is ‘devoid of economic substance’ must be disregarded for tax purposes” and cannot form the basis of a tax deduction. Id. (quoting Lerman, 939 F.2d at 45); accord Wexler, 31 F.3d at 123. In ACM Partnership, the Third Circuit Court of Appeals outlined the framework for the economic substance analysis as follows:

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. Casebeer [v. Comm’r, 909 F.2d 1360, 1363 (9th Cir.1990) ]; accord James v. [Comm’r], 899 F.2d 905, 908-09 (10th Cir.1990); Rose v. [Comm’r], 868 F.2d 851, 854 (6th Cir.1989).

157 F.3d at 247. In applying these principles, the Court reviews the transaction “as a whole, and each step” thereof. Id. at 247 (internal quotation omitted). The IRS asserts Camelot’s COLI VIII plan lacks both objective economic substance and subjective business purpose. Although these inquiries do not constitute two distinct prongs of analysis, the Court will address these component inquiries in turn. However, before doing so, it is useful to summarize the only federal court decision rendered thus far examining the deductibility of policy loan interest under a broad-based leveraged COLI policy, Winn-Dixie, 113 T.C. 254. In that case, the Tax Court applied the economic substance doctrine to disallow interest deductions generated by a broad-based leveraged COLI plan similar in some respects to the plan at issue in this case.

Winn-Dixie had purchased corporate-owned life insurance policies on the lives of approximately 36,000 of its employees and systematically took policy loans against the policy value of the policies to fund the premiums through netting transactions like those used in Camelot’s COLI VIII plan. Because, as in this case, the Winn-Dixie plan was operated in such manner that virtually the entire policy value was encumbered by policy loans and accrued interest on those loans, the net equity in the policies was negligible.

With regard to its objective economic substance analysis, the Tax Court evaluated the profitability of the Winn-Dixie plan in the absence of the policy loan interest deductions by examining the issue illustrations projecting financial performance over the life of the plan. The court then concluded that “without the tax deductions, the plan as designed would produce a negative cash-flow and a negative earnings effect for petitioner in each and every year the plan was in effect.” Id. at 285. Although Winn-Dixie argued that it acquired its COLI policies to help fund employee benefits, in its analysis of the taxpayer’s subjective business purpose, the Tax Court found no contemporaneous evidence to support this argument. See id. at 286. Because there was no evidence Winn-Dixie sought to insure against a “catastrophic event” and the evidence in the record established that interest deductions were the reason Winn-Dixie purchased its COLI policies, the Tax Court concluded that “no nontax purpose was served by the COLI transactions.” Id. at 285-86, 290. The Tax Court then held Winn-Dixie’s COLI program lacked objective economic substance and business purpose other than tax reduction. 48 See id. at 290.

a. Objective Economic Substance of Camelot’s COLI Transaction

In its discussion of the objective aspect of the economic sham analysis, the Third Circuit appellate court in ACM noted that “the courts have examined 'whether the transaction has any practical economic effects other than the creation of income tax losses.’ ” 157 F.3d 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 ‘nontax substance’ because they ‘did not appreciably affect [the taxpayer’s] beneficial interest except to reduce his tax.’ ” Id. (quoting Knetsch, 364 U.S. at 366, 81 S.Ct. 132) (brackets in original). The court concluded that ACM’s transactions “had no effect on [its] net economic position or non-tax business interests,” and that “[t]ax losses such as these, which are purely an artifact of tax accounting methods and which do not correspond to any actual economic losses, do not constitute the type of ‘bona fide’ losses that are deductible under the Internal Revenue Code and regulations.” Id. at 251-52.

The IRS asserts Camelot’s COLI VIII plan lacks objective economic substance because: (1) as in Winn-Dixie, an aggregate economic analysis of the plan’s expected cash flows and earnings results demonstrates that Camelot’s COLI VIII plan could not be reasonably expected to generate a profit absent the interest deductions (hereinafter a “pre-interest deduction profit”); and/or (2) Camelot could not expect to profit from the non-interest deduction related economic components of Camelot’s COLI VIII plan, that is, the inside cash build-up or the death benefit proceeds. Camelot responds: (1) it is improper for the Court to analyze the profitability of the plan without taking into account the tax-favored aspects of life insurance, but that, nonetheless, its COLI VIII plan would have been profitable absent the loan interest deductions; and (2) the inside cash build-up and death benefits components, as well as Camelot’s payment of substantial premiums through $12 million in cash and loans over the first seven policy years and its payment of outstanding policy loans and interest from death benefits proceeds, further demonstrate its COLI VIII plan changed its net economic position and non-tax interests. Additionally, Camelot asserts the economic substance of its COLI VIII plan is demonstrated because there were economic consequences to MBL and Hartford’s reporting for purposes of federal and state taxes of the annual premiums received and policy loan interest earned. The Court will address in turn (1) the pre-interest deduction profit of Camelot’s COLI VIII plan as a whole; (2) the substance of the non-tax deduction components of the plan; and (3) the effect of MBL and Hartford’s reporting of the premiums and policy loan interest on the plan’s economic substance; and (4) what Camelot got from its high premiums and policy loans.

(1) Pre-Interest Deduction Profit

This aspect of the analysis focuses objectively on whether Camelot would have achieved any economic benefit from purchasing its COLI VIII policy plan absent the interest deductions. See Winn-Dixie, 113 T.C. at 260-61. The IRS asserts that, absent the tax deductions taken for policy loan interest, Camelot’s COLI VIII purchase would be a “financial disaster.” D.I. 118, at 31 (Petitioner’s Legal Argument and Evidentiary Objections). The IRS points to several 20- and 40-year COLI VIII plan illustrations that Camelot received prior to and contemporaneous with the purchase of its COLI VIII plan and which project, on an aggregate basis, the expected overall cash flow from the plan. According to these illustrations, the IRS contends, Camelot’s net cash flow was expected to be negative in all years and negative overall without the interest deductions. Camelot points to alternative 81-year illustrations, which it contends prove Camelot could expect a net positive cash flow over the life of the COLI VIII plan, even absent the interest deductions. The IRS counters that the illustrations on which Camelot relies also show its COLI VIII plan to be an economic loser absent the interest deductions. Moreover, the IRS asserts, these illustrations are flawed, and corrected re-illustrations covering the life of the plan show that in the aggregate, Camelot’s COLI VIII plan would generate negative pre-interest deduction cash flows over the life of the plan.

Before turning to an analysis of the cash flow illustrations, the Court will first address Camelot’s argument as to why it is inappropriate to analyze the pre-interest deduction profitability of its COLI VIII plan. Camelot contends, because life insurance is a tax-favored product, the Court should examine the objective economic substance of its COLI transaction only on an after-tax basis. Camelot relies on the following quote from Sacks v. Comm’r, 69 F.3d 982, 991 (9th Cir.1995): “Where a transaction has economic substance, it does not become a sham merely because it is likely to be unprofitable on a pre-tax basis.” 49

Sacks is not dispositive of this case. Sacks involved the allowability of depreciation deductions and investment credits on a sale and leaseback transaction regarding solar energy equipment. Figuring largely in the reasoning of the Sacks court was the fact that both federal and state legislatures had explicitly created the tax credits and accelerated deductions at issue explicitly in order to encourage more investment in solar energy:

Absence of pretax profitability does not show “whether the transaction had economic substance beyond the creation of tax benefits,” Casebeer, 909 F.2d at 1365, where Congress has 'purposely used tax incentives to change investors’ conduct. Congress and the Arizona legislature purposely skewed the neutrality of the tax system even more than the usual tax credits and accelerated depreciation designed to encourage investment in capital goods than would otherwise be made, because they sought to induce people to invest in solar energy.

Id. at 991 (emphasis added). The court further cited the legislative history of the relevant code provisions, id. (“the committee feels that there is a need to encourage the purchase and installation of this equipment” (citations omitted)), in support of its conclusion that, in enacting the targeted tax advantages in question, Congress “intended to induce investments which otherwise would not have been made.” Id. at 992.

In contrast to Sacks, at issue in this case is whether Camelot’s policy loan interest deductions are deductible under the generic, non-targeted I.R.C. § 163(a). Camelot has not pointed to anything to support the notion that Congress, in enacting any tax provision, specifically intended to induce parties to invest in the type of broad-based leveraged COLI VIII plan at issue in this case. At best, Camelot could cite to § 264’s “four out of seven safe harbor.” However, the safe harbor provision is an exception to the general prohibition on policy loan interest deductions with respect to systematic borrowing of the policy value of life insurance contracts. Far from trying to encourage the type of tax-driven, highly-leveraged, broad-based COLI at issue here, the legislative history of the Revenue Act of 1964 states that Congress’ amendments to § 264 limiting the deducti-bility of policy loan interest were aimed at “minimizing the sale of insurance as a tax-saving device.” H.R. REP. NO. 749 (1963), reprinted in 1964 U.S.C.C.A.N. 1313, 1370; S. REP. NO. 830 (1964), reprinted in 1964 U.S.C.C.A.N. 1673, 1750. The Committee reports particularly point out the undesirability of insurance policies that contemplate a systematic plan of borrowing the cash value of policies in such a manner that taxpayers, through the combination of inside build-up and policy loan interest deductions, can actually profit from the transaction. Id. The four out of seven safe harbor, far from being designed to induce the type of tax-driven plan at issue here, is a recognition of “the importance of being able to borrow on insurance policies ... for other than tax saving purposes.” Id. (emphasis added). 50

The Court is unaware of any statute or legislative history indicating Congress has condoned, much less encouraged, policy loan interest deductions of the type and magnitude involved in Camelot’s broad-based COLI VIII plan. To the contrary, 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, e.g., § 264’s general prohibition on deducting policy loan interest incurred as part of a plan to systematically borrow part or all of the increases in the policy values of the insurance, the $50,000 loan limit, and most recently the relevant provisions of HIPA. It follows, the reasoning of Sacks does not apply here. 51

(a) 20- and 40-Year Illustrations Received by Camelot Prior to and Contemporaneous with its COLI Purchase

Camelot received a series of 20- and 40-year cash flow illustrations from Newport prior to its purchase of the MBL COLI VIII policies. All of the illustrations received by Camelot prior to and contemporaneous with its COLI VIII purchase show the same phenomenon: for each and every plan year and on an aggregate basis, over the period of the illustration, cash flows from the plan were projected to be positive with the interest deductions and negative without them. J17; J33; J35/G403; J44; J188.

Although none of the illustrations contain a column showing the pre-interest deduction cash flow, it is a simple calculation. The illustrations show the interest deduction and the post-interest deduction cash flow. The pre-interest deduction cash flow is simply the post-interest deduction cash flow, less the interest deduction. Column 11 of the modified version of one of these illustrations in Appendix A contains this calculation.

For example, the illustration forwarded to Camelot immediately after purchase of the COLI VIII policies (with 1431 insureds and a premium of $10,000 per insured) shows expected aggregate loan interest deductions of $60,001,000 and an after-tax cash flow of $40,639,000. J44 at Bates CAM 004937; Appendix A. The difference, approximately negative $19.3 million, is the pre-interest deduction cash flow. Appendix A. In other words, in the absence of the interest deductions, Camelot’s COLI VIII plan would result in a loss of more than $19 million over the 20-year period illustrated. Similarly, the illustration also reveals the pre-interest deduction cash flows would be negative for each and every policy year. J44; Appendix A. In sum, with the benefit of the policy loan interest deductions, Camelot’s COLI VIII plan was projected to produce large positive cash flows and earnings but that, absent those loan interest deductions, the plans would produce negative cash flows for each and every year and in the aggregate.

Camelot maintains this analysis is flawed because it does not reflect that death benefits are taxfree. Camelot con tends, to put this tax-advantaged life insurance product on a comparable footing with non-tax-advantaged investments, the Court must “gross-up” the COLI policy death benefits to take into account Camelot’s 40 percent tax bracket when doing any pre-interest deduction profitability analysis. Because they are tax exempt, Camelot contends the projected $2,041 billion aggregate death benefits must be grossed up to be equivalent to a stream of taxable income of $3.6 billion. Respondent’s Post-Trial Brief, D.1.123 at 60.

It is inappropriate to “gross-up” Camelot’s death benefits when evaluating the economic substance of Camelot’s transaction for several reasons. The experts who testified at trial made it clear that, although “grossing-up” may be a useful “rule of thumb” to quickly compare the return on some types of tax free investments with the return on taxable investments, grossing up tax-favored income is not a correct financial method to analyze the economic substance of the transaction because a gross-up does not reflect the actual cash flows of an investment. Hoag, Tr. 4600-01; 4764; Puglisi, Tr. 4370. Camelot presented no expert testimony to support its position on the financial significance of grossing up tax-favored income or to rebut the opinions offered by Hoag and Puglisi.

Second, Camelot’s position at the trial was that, in making its purchase decision, it was solely interested in “cash flow” derived from its COLI VIII plan. Rogers, Tr. 639-41. Every illustration that Camelot received (whether before or after its purchase) focused on the ultimate financial consequence of the after-tax cash flows of the plan. J17; J33; J35/G403; J44; J188. The illustrations do not show the pre-tax cash flows of the COLI VIII plan, much less grossed-up pre-tax cash flows. Id. Given Camelot’s stated objective of acquiring this investment plan to obtain a stream of positive and expendable cash flows, the grossing up of the pre-tax cash flows would be totally meaningless. Rogers and Hoag agreed that Camelot could not purchase anything with “grossed-up” cash flows, but could only use actual cash to purchase medical benefits, or anything else. Rogers, Tr. 641; Hoag, Tr. 4601. Finally, there is no evidence in the record that Camelot did any sort of analysis comparing the returns of the COLI VIII plan, grossed-up or not, to that of alternative taxable investments, indicating that, at least at the time, Camelot did not find grossing-up to be particularly useful. For the foregoing reasons, the Court will not “gross-up” the death benefits when evaluating the economic substance of the COLI VIII plan.

In sum, the 20- and 40-year cash flow illustrations received by Camelot immediately prior to and after its purchase of the COLI VIII policies, 52 and reflecting the premium level and number of insureds in the COLI VIII plan, 53 demonstrate that, absent the loan interest deductions, the plan would generate significant negative cash flows for each and every plan year and in the aggregate.

(b) 81-Year Illustrations

Camelot further contends that, even if the profit potential of Camelot’s COLI VIII policy is analyzed on a pre-interest deduction basis and is not “grossed-up,” Camelot’s plan would generate a pre-inter-est deduction profit over the life of the plan. It asserts this distinguishes Winn-Dixie and shows that COLI would not have been a “financial disaster” absent the policy loan interest deductions. Camelot maintains it is inappropriate to examine the 20- and 40-year illustrations discussed above; rather, it urges, particularly in light of the young age of its workforce and the larger death benefits Camelot anticipated receiving in later plan years, the Court should look to plan illustrations covering the projected 81-year fife of the plan. It then relies on a set of 81-year issue illustrations, produced by IAS and forwarded to Newport by memorandum dated May 1,1990, 54 to show that its COLI VIII policy had positive cash flows and earnings even absent the tax deductions for loan interest. J42/43. (“May 1, 1990 illustrations”) (modified version attached at Appendix C).

The specific illustrations that Camelot relies on are labeled as Scenario 3, which consist of a cash flow illustration and an earnings illustration. 55 The Scenario 3 illustration is nearly identical for the first 20 years to the 20-year illustration, discussed above, that Campisi forwarded to Rogers on May 3, 1990. Compare J44 with J42, Scenario 3. The May 1, 1990 illustration shows that without the policy loan interest deductions, cash flows would be negative each year until the 53rd year of the plan. However, the illustration demonstrates a projected after-tax aggregate cash flow of $383 million over the life of the plan. Appendix C. Over this period, the aggregate cash flow from the policy loan interest deductions was projected to total approximately $180 million. Appendix C. Deducting the latter from the former, the projection shows an aggregate pre-interest deduction profit of approximately $200 million. Appendix C. Because this illustration shows an aggregate pre-interest deduction profit, Camelot argues, net death benefits were “clearly the dominant element,” demonstrating that Camelot’s COLI VIII plan had economic substance. 56

The IRS counters, first, the Court should not rely on the May 1,1990 illustration because it is fatally flawed and inaccurate. Second, the IRS argues, even using the fatally-flawed illustration, a present value analysis demonstrates Camelot’s COLI investment would have been financially illogical absent the tax benefits of the policy loan interest deductions.

As to the IRS’s first argument, the Court agrees the May 1, 1990 issue illustration does not reflect the objective economic substance of the COLI VIII plan that Camelot purchased since, as Campisi acknowledged, it was flawed and did not accurately reflect the plan’s intended consequences. 57 Campisi, Tr. 1531-36. The illustration projected negative after-tax cash flows in plan years 47 to 53 ranging from approximately $3 million to $22 million, totaling more than $97 million. J42; Campisi, Tr. 1535-36. These large negative cash flows were at odds with both the design objective of the COLI VIII product to achieve positive cash flows in every plan year and Camelot’s professed business purpose for purchasing the plan, i.e., to obtain a stream of positive cash flows that could be used to offset medical benefit expenses. G332, Bates M005170; Eisen-berg, Tr. 2426-29. Given Camelot’s stated business purpose to obtain positive cash flows from the COLI VIII plan, and its minimum cash outlay strategy, it is very doubtful it actually intended to invest $97 million in cash in the program in these later plan years. Camelot’s desire not to make a substantial cash investment in its COLI VIII plan is evidenced by every sales illustration and later plan reissue illustration, not one of which showed Camelot being required to make a relatively low substantial cash investment in the plan. J17; J18; J33; J35/G403; J44; J188.

The negative cash flows in the May 1, 1990, issue illustration appear to have been produced by a purported flaw in IAS’s illustration program that produced negative cash flows in the later years of plans in which the insured employee population was relatively young. A warning in an internal IAS memorandum addressing illustration procedures for the COLI product indicates IAS was aware of this problem: “Before sending the [issue] illustrations out, Schedule A cash flows should be checked to see if they are negative. If a large proportion of people are under age 30, carefully check after year 30.” G315, Bates 015166. When IAS forwarded the May 1, 1990, 81-year illustration to Newport, it informed Peter Cahall, Campisi’s partner, of the unintended large negative cash flows illustrated in plan years 47 through 53. G938; Campisi, Tr. 1535. Because of these large unintended negative cash flows, Campisi decided not to give the illustration to Camelot at that time. 58 Campisi, Tr. 1535-36. On these facts, it is difficult to accept Camelot’s argument that this flawed illustration should form the basis for a finding that Camelot’s plan had economic substance.

Moreover, after Camelot purchased the COLI VIII plan, IAS corrected the problem in the illustration program and generated new illustrations for Camelot’s COLI VIII plan that substantially reduced these negative after-tax cash flows. The huge negative cash flows for years 47 through 53 were largely corrected by the second policy year and completely eliminated in later policy years. Campisi, Tr. 1707-08; see also J77, at 5-6; J110, at 5; Jill, at 8-11; J125, at 8-11; J129, at 8-11. For example, Newport’s February 1992 Plan Anniversary Review of Camelot’s COLI VIII plan included a “re-illustration” of the expected cash flows in which the projected negative after-tax cash flows totaling $97 million in plan years 47 through 53 were reduced. J77 at 5-6. The re-illustration projects negative after-tax cash flows in plan years 49 through 53, totaling approximately $15.7 million, a significant reduction from the May 1, 1990 illustration. J77. Interest deduction tax credits are projected to exceed after-tax cash flows for every year up to year 55, after which point after-tax cash flows exceed interest deductions for each year up to plan termination. J77, at 5-6. However, total after-tax cash flow over the life of the plan is estimated to be $167.7 million; whereas total interest deductions over the life of the plan are projected to be $177 million. Thus, under this set of “corrected” projections, absent the tax deductions for policy loan interest, the aggregate cash flow would be negative approximately $9 million over the life of the plan. Moreover, while the Court need not rely on them for its conclusion, an analysis of each subsequent plan anniversary re-illustration of Camelot’s COLI policy cash flows reveals that, with these corrected re-illustrations, the pre-interest deduction profits that Camelot points to as evidence of the economic substance of its COLI VIII plan evaporate. 59

Even if the flawed May 1, 1990 illustration were used to evaluate economic substance, the Court is not persuaded it demonstrates that Camelot’s COLI VIII plan has substance apart from the tax benefits. As previously stated, the illustration shows pre-interest deduction cash flows do not turn positive until the 53rd year of the plan; that is, without the tax benefits from the interest deductions, Camelot could not have expected to receive positive cash flows that could be used to offset employee benefit costs until 2043.

Where Camelot has to wait more than half a century for pre-interest deduction cash flows to turn positive, evaluating the present value of the cash flow is the proper way to evaluate the profitability of the May 1, 1990, issue illustrations. See ACM Partnership, 157 F.3d at 259 (“In transactions that are designed to yield deferred rather than immediate returns, present value adjustments are, as the courts have recognized, an appropriate means of assessing the transaction’s actual and antici pated economic effects.”)- 60 Camelot, by doing a present value analysis, implicitly acknowledges that mode of analysis is appropriate in this case, and argues the 81-year illustration shows the present value of the aggregate cash flows and earnings, whether grossed up or not, are positive. See D.I. 123, at 65-67.

Hoag testified that the present value to Camelot of the positive projected pre-in-terest deduction cash flows in the later plan years was negative at the 8% discount rate used to calculate present values in the issue illustrations themselves. Hoag, Tr. 4598-4600, R238/G740, at 14. Hoag also maintains that the present value of the projected pre-interest deduction cash flows was negative at any reasonable discount rate. 61 See id. Camelot made no attempt to present expert testimony or any other evidence rebutting Hoag’s testimony. Instead, in its post-trial brief, Camelot attempts to supplement the record by providing calculations demonstrating the present value of pre-interest deduction cash flows and earnings would be positive at discount rates of less than or equal to 3.6584 % and 4.7954%, respectively. 62 D.I. 123, at 66. Camelot asserts “[t]hese discount rates are reasonable given the fact that net death benefits are not subject to tax and that the guaranteed crediting rate in the COLI VIII policies was 4 percent.” Id. at 67.

Camelot’s argument fails for two reasons. First, the present value calculations that Camelot included in its post-trial briefs and appendices are not part of the record evidence presented at trial. Moreover, although Camelot submits that discount rates of 3.6584% and 4.7954% would be “reasonable,” it fails to provide any support for this proposition in the eviden-tiary record, financial literature, or case law. What evidence there is in the record suggests that these are not “reasonable” discount rates. The discount rates presented in Camelot’s Post-Trial Brief are substantially below the 18-20% rate of return that Camelot claims to have been receiving on its capital investments, and the 8% rate used to illustrate present values in the COLI VIII illustrations themselves, see, e.g., J42/43 and the 8.61% rate of interest that it would have earned had it instead invested in risk free, thirty-year maturity United States Treasury Bonds as of February 16, 1900. See FEDERAL RESERVE BULLETIN A24 (May 1990); reprinted in Petitioner’s Reply Brief, Appendix E. 63 It has been long established that, in calculating the present value of a lost stream of income, the minimum discount rate should be the rate of interest earned on “the best and safest investments.” Chesapeake & Ohio Ry. Co. v. Kelly, 241 U.S. 485, 490-91, 36 S.Ct. 630, 60 L.Ed. 1117 (1916); see also Jones & Laughlin Steel Corp. v. Pfeifer, 462 U.S. 523, 537, 103 S.Ct. 2541, 76 L.Ed.2d 768 (1983). The thirty-year term Treasury Bond rate qualifies as one of the best and safest investments. Camelot did not brief as of what date the discount rate should be applied nor did it place on the record the rate on any particular date. As Hoag testified and Camelot’s own calculations confirm, the present value of the pre-inter-est deduction cash flow projected in the May 1, 1990, issue illustrations at an 8 percent discount rate is negative. Therefore, the Court finds that Camelot has failed to present sufficient evidence that in the aggregate using present value analysis its COLI VIII plan would produce a net economic benefit in the absence of the interest deductions.

In sum, the Court concludes the May 1, 1990 81-year illustration relied on by Camelot is insufficient evidence to support a finding that Camelot’s COLI VIII plan changed its net economic position apart from the claimed tax deductions. First, and the most likely reason Camelot was not provided with the illustration until long after its COLI VIII plan had been in effect, is that the illustration was considered by IAS and Newport to be fundamentally flawed due to huge negative cash flows in plan years 47 through 53 that were at odds with the plan design. Second, in the unlikely event the Court were to credit the validity of the illustration, Camelot did not adduce sufficient evidence to prove the COLI VIII plan would provide a reasonable return in the absence of the loan interest deductions.

Based on an aggregate analysis of pre-interest deduction cash flows, the Court concludes Camelot failed to establish that its COLI VIII plan had objective economic substance.

(2) Substance of Non-tax Components of Camelot’s COLI VIII Policy

An analysis of the economic benefits of the non-interest deduction related components of Camelot’s COLI VIII plan further support the conclusion that the transaction lacked objective economic substance. As set forth supra, section II.A.1. and 2., whole life or universal life insurance policies have two components from which a policyholder may derive economic benefit: tax-deferred inside build-up of policy cash value and tax-free death benefits. The evidence demonstrates Camelot could not expect to derive material economic gain from either of these components.

(a) Inside Cash Value Build-Up

As previously discussed, supra, section H.D.4., it is well established in the record that Camelot’s COLI VIII plan was designed and operated so as to attain zero net equity with respect to each of its policies at the end of each policy year. 64 Zero net equity in each policy year was an important design feature for plan performance, since unborrowed equity left in the policy generally had a negative impact on cash flows to the policyholder. G889, Bates 001890; see also Eisenberg, Tr. 2356-64; Kristie Sayre, Tr. 3998^1002. Consistent with this design, Camelot’s COLI VIII plan was administered using a computer system programmed to ensure that the amount of net equity in each policy at the end of a policy year would not exceed one penny. Eisenberg, Tr. 2225-26, 2518-22; Kristie Sayre, Tr. 4002-09; G844, Bates 5583; G848, at 2; G949, Bates 015150.

The effect of maintaining a zero net equity at the end of every policy year was that Camelot’s COLI VIII plan did not build up or accumulate any significant unencumbered policy value over the term of the plan. 65 This being so, Camelot could not expect to derive any economic gain from inside build-up. With a net equity of zero, should Camelot surrender the poli cies, it would walk away with essentially nothing. Moreover, the purported inside build-up would not contribute to greater death benefits because virtually the entire policy value was encumbered by policy loans and accrued interest, which would reduce the death benefit by that entire amount. Therefore, the accumulated cash value component of Camelot’s policies was an economic neutral in Camelot’s calculus of profit.

MBL recognized that this zero net equity feature was a significant indicator of the COLI VIII plan’s lack of economic substance. When a tax lawyer for a COLI broker suggested to MBL’s Wendell Bos-sen that it would be very difficult to convince a court of the economic substance of the COLI arrangement given the “uninterrupted string of zeros” in the net equity column of the COLI VIII product illustrations, Bossen recommended to McCamish that the net equity column be eliminated “since anything we can do to remove self-made traps in the illustration would be helpful.” G189; see also Bossen Tr.2000-01; 2019-21, 2110-12; Eisenberg, Tr. 2286-87. As evidenced by Camelot’s COLI sales and issue illustrations, which do not have such a column, McCamish acted on Bossen’s advice. See, e.g., J33, J35, J44.

Equally telling, Camelot justifies its choice of the highest possible loaned crediting rate by asserting it would contribute to a higher inside cash build-up which would not be taxed. Campisi, Tr. 1325-26; Eisenberg, Tr. 2260, 2304-06. However, as testified, the fixed spread between the loaned crediting rate and the loan interest rate effectively made the election of one the election of both. Eisenberg, Tr. 2324-26; Campisi, Tr. 1332-33; Kristie Sayre, Tr. 3964-65, R314/G738 at 24, 32; G56 at Bates NJ650. Moreover, because of the fixed 1% spread between the loan interest rate and the loaned crediting rate, any greater inside build-up would be encumbered by even larger accrued loan interest, so that Camelot would obtain no economic advantage from the higher inside build-up. Campisi, Tr. 1338-39; McGill, Tr. 3721-22; Hoag, Tr. 4614-17. The only advantage Camelot achieved by electing the highest loaned crediting rate/loan interest rate combination was the ability to take larger interest deductions against taxable income in computing its federal income tax. 66 Ei-senberg, Tr. 2400-03; Campisi, Tr. 1363-64; Kristie Sayre, Tr. 3971-80; R314/G738 at 33-35; Puglisi, Tr. 4364-74, R188/G739 at 14-16; Hoag, Tr. 4577-80, R238/G740 at 12-14; G117 at Bates 001675; G967; G968; G969. Thus, the inescapable inference is that Camelot’s choice of the highest possible crediting rate was really the means to attain the higher loan interest rates 67 and is just a further indication that the COLI VIII plan lacked economic substance apart from the interest deduction,

(b) Tax Free Death Benefits

Camelot asserts its COLI VIII plan has economic substance because it would receive significant tax free death benefits. However, the preponderance of the evidence shows that the COLI VIII plan was designed to be mortality neutral. That is, the cumulative COI 68 charge paid by Cam elot was anticipated to equal the cumulative death benefit that would be distributed to Camelot, with the exception of a profit margin to the insurance company of 20% of the COI charge in the first plan year, 10% in the second plan year, and 2% thereafter. Eisenberg, Tr. 2620, 2663.

In traditional life insurance, the policyholder’s beneficiary will profit if the insured dies prematurely. See Winn-Dixie, 113 T.C. at 284 (“We recognize that one of the normal benefits of life insurance is the death benefit to be received if the insured dies before the insured’s actuarially determined life expectancy.”). In contrast, Camelot’s COLI VIII policy was designed so that neither Camelot nor MBL was expected to profit from mortality, other than MBL’s small built-in mortality profit margin; rather, the COI charges were set at a level that would reflect projected policyholders’ actual mortality experience, thereby ensuring that mortality would play a neutral role in the product’s financial performance. Eisenberg, Tr. 2562-68, 2623-24; Van Etten, Tr. 2755-57, 3036-44, 3046-48; Sayre, Tr. 4106-07; Hoag, Tr. 4611-14; G17, Bates 000257; G82, Bates M & R 001670; G136, Bates M & R 009543; G330, Bates M & R 007096. As stated in a M & R Executive Summary describing the COLI VIII product:

Since the sale is tax driven, mortality should play a neutral role in perfor■mance. 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.

G527, Bates 027021 (emphasis added). Amother M & R product description defined the COLI VIII product’s mortality neutrality feature as a “partnership” between the policyholder and insurance carrier to ensure that neither party profited or sustained losses based on actual mortality experience:

The mortality risks [of broad-based COLI] are more similar to group insurance than traditional individual insurance. 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.

G852, Bates 027016 (emphasis added). Stated differently, mortality experience was intended to be a neutral factor in the financial performance of the COLI VIII product. Kristie Sayre, Tr. 4107-09; Hoag, Tr. 4611-14.

It follows that in contrast to traditional whole or universal life insurance policies in which the policyholder’s beneficiary stands to profit from mortality in the event of premature death, Camelot could not expect to profit from mortality over the duration of its plan or even in any year of its plan. This was evidenced at trial by Rogers, who mathematically demonstrated that the COLI death benefits were simply insufficient to provide Camelot with a profit, even over 30 years. Rogers, Tr. 445-46; J188; G993.

Camelot correctly points out that, despite the evidence its COLI VIII plan was designed to be mortality neutral, it did not perform in that manner. From the plan inception through December of 1998, Camelot received death benefits plus mortality dividends in an amount $1.3 million higher than it paid in cost of insurance charges during that period. Van Etten, Tr. 2761-63; R572; G964. Camelot asserts this mortality gain is evidence that Camelot bene-fítted from mortality risk transfer to MBL and Hartford. Van Etten, Tr. 2760-64. Such risk shifting and risk transfer are the essence of life insurance. See Helvering v. Le Gierse, 312 U.S. 531, 539, 61 S.Ct. 646, 85 L.Ed. 996 (1941).

While Camelot’s mortality experience was more favorable than expected for the first seven years of its plan, this windfall is hardly sufficient to imbue its COLI VIII plan with economic substance. First, even the additional $1.3 million was insufficient to make Camelot’s COLI VIII plan profitable without the interest deductions during this time period. See Appendix B. Further, the part of Camelot’s “mortality gain” attributable to mortality dividends does not reflect risk shifting, but instead is a function of how the COLI VIII plans were administered by MBL and Hartford in an effort to ensure mortality neutrality on all of their COLI VIII plans. Although Camelot experienced better than projected mortality in the early plan years, in the sense that it had more employee deaths than expected, most COLI policyholders had the opposite experience; that is, MBL experienced considerable mortality profits on its COLI VIII policies overall. Eisen-berg, Tr. 2248-49, 2565-66; Bossen, Tr.2078-81; G275, Bates MR4976; G281, Bates ICMG 876; G486, Bates 018373. MBL came under pressure from brokers and policyholders, angered because the COLI VIII policies were not living up to their mortality-neutrality billing. Bossen, Tr.2076-77, 2081-88; Eisenberg, Tr. 2576; Van Etten, Tr. 3049-53; G276; G373, Bates 020585; G433; G437, Bates 007228; G464. In response, MBL began declaring mortality dividends. Bossen, Tr.2088-90; Eisenberg, Tr. 22484,9, 2573-76; Van Et-ten, Tr. 2758-59, 3063-68;G467; G474, at 2; J67. Although the dividends initially were based on the “pooled” experience of all COLI policyholders, beginning in 1993, Hartford began moving its larger COLI plans to an “experience rated” approach, that is calculating mortality dividends based on the specific experience of individual COLI plans, to enable the COLI VIII products to perform and produce financial results that were closer to plan illustrations. Bossen, Tr.2092-93; Eisenberg, Tr. 2571-73; Van Etten, Tr. 3070-78; G421, Bates 014645; G597, Bates ICMG 834; G619, Bates MBL 392-93; G629. Camelot benefitted from the receipt of pooled dividends in 1991, 1992, and 1993 and a combination of pooled and experience rated dividends in 1995. Bossen, Tr. 1783; Eisenberg, Tr. 2249, 2761-62; G963; G964. To the extent Camelot’s “mortality gains” are sourced by mortality dividends, 69 they do not reflect risk shifting but are merely a function of how the plans were administered to try to reflect projected plan performance.

Moreover, as a result of the move to the experience-rated approach that involves not only dividends to policyholders with worse (fewer deaths) than projected experience, but also adding surcharges to policyholder COI charges so that Hartford could recoup its mortality losses from policyholders, like Camelot, who have better than projected mortality experience, it is highly unlikely that Camelot will experience another mortality windfall. 70 The whole point of moving toward an experience-rated approach was to better effectuate the mortality-neutral feature of COLI and thereby ensure predictable plan performance, as explained in a 1995 letter from Campisi to Camelot announcing the extension of the experience-rated approach to Camelot:

The new mortality mechanism will result in a much closer match between the expected cash flow from projected death benefits and claims that are actually paid. This will minimize any volatility or variation in the cash flows and corporate earnings which are expected in each year of the plan.

J122, Bates 00001278; see also Eisenberg, Tr. 2571-73; Van Etten, Tr. 3070-78; Bos-sen, Tr.2092-93; G421, Bates 014645; G597, Bates ICMG 834; G619, Bates MBL 392-93; G629. Although Hartford may never recoup Camelot’s mortality windfall through increased COI charges, this is a function of a policy contract provision capping maximum COI charges, not a function of risk shifting. 71 Camelot has received only one experience rated dividend in the amount of $293 since 1995. J125; J130; J136; J142.

In conclusion, Camelot could not expect to derive substantial economic gain from either the cash build-up or mortality components of its COLI VIII plan.

(3) Substance of MBL/Hartford’s Reporting of Transaction

An analysis of objective economic substance includes an assessment of the economic consequences of the transaction to non-taxpayer parties involved in the transaction. See Frank Lyon Co. v. United States, 435 U.S. 561, 580, 583, 98 S.Ct. 1291, 55 L.Ed.2d 550 (1978) (finding sale-leaseback transaction had economic substance, relying in part on the “absence of any differential in tax rates and of special tax circumstances for one of the parties” and on the fact that “[n]o deduction was created that is not either matched by an item of income or that would not have been available to one of the parties if the transaction had been arranged differently”); cf. ACM, 157 F.3d at 252 n. 40 (“The participation of a foreign partner that was impervious to tax considerations and that claimed most of the reported gains while allocating to Colgate virtually all of the losses allowed Colgate as ACM’s major U.S. partner to reap the benefits of the tax losses without sustaining the burdens of the offsetting tax gains.”). Camelot contends that MBL and Hartford’s reporting of its COLI transactions establishes the Camelot plan had economic substance.

MBL and Hartford reported premiums due under the COLI contracts, for purposes of state premium taxes 72 and, by the second year of the Camelot policy, the federal DAC tax. 73 DesRochers, Tr.1980, 2008-09; Bossen, Tr. 2127-28; Van Etten, Tr. 2724-26, 2766-68. Policy loan interest was also reported for state and federal tax purposes as income by the life insurance companies. DesRochers, Tr.1984. Therefore, Camelot contends, Camelot’s COLI VIII policies had an economic effect on the life insurance companies.

With regard to MBL/Hartford’s reporting of the COLI premiums for state and federal tax purposes, such reporting did not alter the insurance company’s net economic position. This was so because the costs of state premium and federal DAC taxes were included in the actual expenses of the COLI VIII policies, Van Etten, Tr. at 2724, which expenses were covered by Camelot’s cash payments. Thus, whether MBL/Hartford reported the premiums for state and federal DAC purposes, it did not have economic consequences for MBL/Hartford because Camelot’s cash payments covered its share of these expenses, 74 which are essentially transaction costs of the plan. Moreover, as discussed, supra, section III.A.2.d., the premiums reported for purposes of state premium taxes were reduced by the amount of the huge loading dividends in policy years four through seven.

The reporting of the policy loan interest as taxable income is more problematic. If there was evidence that MBL/Hartford reported and paid income tax on the entire policy loan interest from the broad-based leveraged COLI VIII policies, this would be strong evidence the transaction was not a sham in substance. See Frank Lyon Co., 435 U.S. at 580, 98 S.Ct. 1291 (finding it relevant that “[n]o deduction was created that is not ... matched by an item of income”). The Court is troubled, however, by the incomplete picture presented both by the testimony at trial and by the declarations of the Hartford employees proffered by Camelot. 75 There is simply nothing in the record to inform the Court whether, as it strongly suspects, MBL/Hartford also deducted as an expense the interest that it credited to Camelot as inside build-up on the loaned portion of Camelot’s policy value. If the latter is the ease, MBL/Hartford would not be reporting interest income that matched Camelot’s’ deduction, but in reality would only be paying income tax on its profit, that is, the one percent spread between the policy loan interest rate and the loaned crediting rate. Because Camelot has neither provided nor proffered evidence to demonstrate that MBL or Hartford has paid federal taxes on an amount close to or equivalent to the policy loan interest deductions taken by Camelot, it has not sufficiently demonstrated that MBL/Hartford’s reporting of elements of the COLI VIII policies provided the necessary economic substance.

(4) What Camelot Got For its High Premiums and Policy Loans

The amount of actual cash paid by Camelot to MBL in any given year equated exactly with the sum of the COI charges, plan expenses (e.g., broker commissions, fees to McCamish and/or entities he controlled, state premium taxes, DAC taxes), and MBL’s profit for this policy, less a small amount of interest credited by MBL on the policy’s decreasing minimal net equity during the course of the policy year. 76 Kristie Sayre, G738, at 50. While the net cost of Camelot’s death benefit coverage for a 28-year-old male employee during its fifth policy year was $151.25 (COI of $238.68 less a mortality dividend of $84.43), Camelot’s fifth year cash payment to MBL for this same policy was a $1,207.86. Kristie Sayre, G738, at 52.

But, if the cash Camelot paid was sufficient to pay the COI charges and plan-related costs and profits to the insurance carrier, what was the rest of Camelot’s high premium being used for? 77 As previously discussed, Camelot was not paying this amount to build net equity in its policy, for at the end of each policy year, there was none. Because Camelot’s small (relative to premiums) cash payments covered the COI charges, the high premiums were not necessary to fuel the future stream of death benefits. The only remaining utility of the high premiums was to create high policy value to serve as collateral for maximum policy loans, which fueled the interest deduction tax benefit, the only other remaining element of value in the COLI VIII plan. Put differently, what Camelot was getting for borrowing and paying interest, the net cost of which was 1%, was the ability to take the large interest deductions which generated the positive cash flow in its COLI policy. 78 Stated succinctly and bluntly, Camelot’s COLI policy loans had no practical economic effect other than to generate income tax deductions.

b. Subjective Aspects of Economic Sham Analysis

With regard to the subjective or business purpose of the transaction, the Court must determine whether the transaction was “entered into without expectation of economic profit and [with] no purpose beyond creating tax deductions.” ACM, 157 F.3d at 253 (quoting Wexler, 31 F.3d at 123-24) (brackets in original). Camelot first argues it had a bona fide business purpose to purchase the COLI VIII plan, in that it sought to recover and provide for the increasing costs of employee medical benefits. D.I. 123 at 58. Second, Camelot asserts the evidence shows it had a reasonable expectation of profitability over the life of the transaction, even absent the tax deductions. 79

(1) Business Purpose

Camelot asserts it purchased the COLI VIII plan for the legitimate business purpose of offsetting in part the costs of its employee medical benefits program. Camelot presented evidence at trial that attracting and retaining quality employees was critical to the success of Camelot’s business. See Rl; Rogers, Tr. 153; Knoll, Tr. 736. In order to accomplish this objective Camelot developed the “Camelot Associate Program,” an employee benefit program offered to all employees who worked a minimum of 20 hours per week that included health and dental coverage in addition to other benefits. See J172; Rogers, Tr. 152, 155, 162-63. Similar to other employers, Camelot experienced escalating health care costs during the late 1980s, which were a major concern to the company. See J6; Jll; R325-28; Rogers, Tr. 159, 161, 165-66, 188; Knoll, Tr. 737; Bald, Tr. 855-56. Moreover, the number of Camelot employees eligible for health benefits more than doubled, from 600 to 1300 associates, between 1987 and 1989. See J6; Stip. ¶ 25. Some documents regarding Camelot’s COLI purchase reflect that Camelot intended to use COLI cash flows to help offset the costs of employee health insurance. See, e.g., J7; J182.

The IRS asserts that Camelot’s professed business purpose for purchasing the COLI VIII product was not genuine based on evidence in the record that, at various points in time, Camelot provided inconsistent explanations as to why it purchased the policies. The IRS further points to the absence of explicit links between discussions of employee medical costs and the COLI VIII plan during Camelot Board meetings, as well as the apparent ignorance of Scott Sterling, Camelot’s Director of Benefits and Risk Management, regarding the COLI VIII plan. Additionally, the government points out that Camelot’s COLI death proceeds were not segregated for use in funding employee benefit obligations, but rather were used for general corporate purposes.

The Court need not resolve whether Camelot’s asserted purpose for the COLI cash flows is genuine because, even if Camelot’s purpose for purchasing the COLI VIII plan was to provide funds for employee medical benefits, such a purpose, standing in isolation, is not sufficient to imbue what is otherwise an economically empty transaction with economic substance. As the Tax Court aptly observed in Wimv-Dixie:

Even if we were to accept Mr. McCook’s [Winn-Dixie’s financial vice president] testimony that he intended to use tax savings to fund Winn-Flex [Winn-Dix-ie’s employee benefits program], that would not cause the COLI plan to have economic substance. If this were sufficient to breathe substance into a transaction whose only purpose was to reduce taxes, every sham tax-shelter device might succeed. Petitioner’s benefit from the COLI plan was dependent on the projected interest and fee deductions that would offset income from petitioner’s normal operations. The possibility that such tax benefits could be used as a general source for petitioner’s [employee benefits] obligations (or any other business purpose) does not alter the fact that the COLI plan itself had only one function and that was to generate tax deductions which were to be used to offset income from its business and thereby reduce petitioner’s income tax liabilities in each year.

113 T.C. at 287-88.

The Court must focus on the purpose of the underlying transaction at issue here, not what Camelot intended to do with the proceeds. 80 Camelot’s purpose in purchasing broad-based leveraged COLI was to acquire an investment vehicle to generate positive cash flows. See, e.g., Rogers, Tr. 639-40. As applied in the context of this case, the test for business purpose is whether the underlying transaction has a purpose beyond creating tax deductions. See ACM, 157 F.3d at 253-54. What the cash flows derived from this transaction were to be used for is irrelevant for the subjective prong of the economic substance analysis. Whether Camelot intended to apply COLI proceeds to the sympathetic purpose of funding employee medical benefits or for any other corporate purpose does not “breathe substance” into a transaction whose purpose is to generate positive cash flows and earnings solely through tax deductions. Winn-Dixie, 113 T.C. at 287-88.

The preponderance of the evidence convincingly demonstrates that policy loan interest tax deductions were the predominant consideration in Camelot’s decision to purchase broad-based leveraged COLI as well as the selection of premium level and interest rates. The Newport Group’s communications with Camelot regarding broad-based COLI highlight that the primary function of the plan is to fuel policy loan interest deductions. When Campisi first wrote to Camelot in the July 6, 1987 letter about the possibility of purchasing the COLI VIII product, he described it, not as a life -insurance policy, but as an “Insured Investment Contract,” emphasizing that “[t]he key factor is being able to absorb the interest deductions.” J7/G934, at Bates 000062. The prominence to the plan of the interest deductions was further emphasized in the first attachment to the letter, a chart entitled “Tax Deduction Per Employee,” illustrating the nominal values of the interest deduction per employee, as well as the present value of the deduction discounted at 8 percent, over a twenty year period. Id., Bates 000063. Further, there was enclosed a chart showing how higher interest rates would fuel higher interest deductions which would generate larger cash flows without requiring an additional net cash outlay by Camelot. Id., Bates 000071.

In later correspondence, Newport further suggested that it would “tailor the size of the program to best fit Camelot’s taxable income expectations.” J23. Each of Camelot’s COLI sales illustrations demonstrates clearly how the tax savings from the policy loan interest deductions would produce positive cash flow in every policy year and that, absent those deductions, cash flows would be negative. See J17, J33, J35/G403, J44,J188. In a March 26, 1990, letter concerning alternative premium amounts, Newport suggested a higher premium “if you feel comfortable with Camelot’s ability to absorb the higher tax benefit.” J33. Newport also told Camelot that spreading the premium across more employees would allow Camelot to use “as much of the tax benefit over $2.5 million as can be absorbed in years 8 and after.” J25.

The principal decision makers at Camelot understood performance of the COLI VIII plan was dependent on the interest tax deductions. Rogers knew that Camelot had to have sufficient taxable income to take full advantage of the policy loan interest deductions in order to realize the projected tax-driven positive cash flows from the COLI VIII product. Rogers, Tr. 169, 406, 407, 409, 410. Additionally, Rogers knew he could increase the magnitude of the cash flows by increasing the magnitude of the interest deductions through higher premiums and higher interest rates. Rogers, Tr. 447-49, 450-51, 487. Indeed, at Rogers request, Lee Ann Thorn, Caihelot’s tax director, projected the corporation’s future taxable income to ensure there was enough income to absorb the interest deduction. Thorn, Tr. 668-69, 696, 699. By contrast, although Camelot asserts its purchased the COLI VIII policies to provide funds to pay future employee medical benefits costs, no evidence was adduced at trial reflecting any contemporaneous anal-yses performed by Camelot comparing the level of projected medical costs with antici pated COLI cash flows. 81 Rogers, Tr. 263-64; Thorn, Tr. 715.

That the interest deduction was the critical feature of the COLI VIII plan to Camelot is also evidenced by the parties’ focus on potential tax legislation. Camelot rushed to bind coverage of the COLI VIII plan by February 20, 1990, the day before Congressional hearings on potential COLI legislation were to begin, with the idea that the policies could be backdated to the date of the application and thereby be grandfathered should any adverse tax legislation be enacted. See J22/J23, Bates CAM004824; J25, Bates S03-01444; Rogers, Tr. 484-87; Campisi, Tr. 1195-97. Moreover, Camelot was so concerned about the availability of the COLI interest deduction that it sought and obtained a “honeymoon letter” from MBL promising a refund of its investment in the event of passage of adverse tax legislation prior to December 31, 1990. J26, Bates CAM004637-38; Stip. ¶ 55; Rogers Tr. 236, 243-44, 493-94, 496; Campisi, Tr. 1197-98,1496; J27; J28; G412.

Additionally, Camelot gave considerable attention to strategies for terminating or unwinding the policies should Congress enact adverse tax legislation. See J25. In a letter by Campisi dated December 5, 1989, responding to a question raised by Rogers regarding Camelot’s ability to terminate the COLI VIII plan, Campisi stated plan termination only made sense if a policyholder could not use the interest deduction or if Congress eliminated the deduction:

Of course, the only reason one would ever wish to terminate a COLI program would be if the tax benefits of the interest deductions could not be utilized over a period of years. To our knowledge, this could only happen for two reasons. First, the company did not generate sufficient taxable income over several years. If this were the case, termination through surrender would also become a viable option because the taxable gain created would not produce a tax liability.

The second reason would be if Congress changed the taxation of life insurance and applied it retroactively to policies purchased prior to the effective date of the legislation.

J15, Bates 000461. Attached to Campisi’s letter were illustrations showing that, if the policy loan interest deductions were eliminated or the policyholder unwound its plan by stopping premiums and holding the policies in nonforfeiture status, the COLI VIII plan would produce negative after-tax cash flows. See J15; see also Rogers, Tr. 423-24, 427; Campisi, Tr. 1405-12.

Moreover, the risks that concerned Camelot with respect to the purchase of the COLI VIII policies all related to the policy loan interest deductions. J25; see also Rogers, Tr. 245-47, 501, 507. As stated in one of Camelot’s internal documents: “The two most likely reasons for freezing the program are a retroactive tax law change or the failure of Camelot to generate sufficient earnings to absorb the interest deductions created by the plan.” J25. That document summarized Camelot’s risks upon purchasing the COLI VIII policies this way: “1) A retroactive tax law change[;] 2) Camelot’s failure to generate taxable income over several years in a row[;] and 3) IRS attack [.]” J25 (emphasis added). All of Camelot’s articulated risks related to loss of the benefit of the loan interest deductions.

Further, once the 1996 HIPA legislation definitively eliminated COLI VIII’s policy loan interest deductions, Camelot quickly instructed Hartford to stop billing it for annual premiums and to allow the policies to function as paid-up policies for a reduced amount of death benefit coverage. See G1051; G1078. That reduction in death benefits (approximately $30,000,000), which was accompanied by a partial withdrawal of policy value in the same amount, forced all of Camelot’s taxable policy value gain out of its COLI VIII policies. 82 One reason for the massive partial withdrawal was to utilize net operating loss carry forwards while Camelot was in Chapter 11. Rogers, Tr. 363. Moreover, notwithstanding the fact that it continued to have employee medical cost obligations, there is nothing in the record by way of testimony or documentation that Camelot was concerned with this substantial reduction in its COLI death benefits inherent in this force-out transaction.

In addition to the rationale articulated by Rogers, a strong inference can be drawn that, without the policy interest deductions, the COLI program was no longer viable. Van Etten, Tr. 2845-46; Kristie Sayre, Tr. 4062-64, 4091-93; G1051; G1078. Support for this inference is provided by Eisenberg:

The Court: Have any COLI VIII policies been sold after passage of the 1996 legislation?

Mr. Eisenberg: I am not aware of any. The Court: And you wouldn’t expect any to be sold?

Mr. Eisenberg: I would not expect a COLI VIII to be sold after that date.

Eisenberg, Tr. 2670. The reason is obvious: The 1996 HIPA legislation eliminated the deductibility of interest on broad-based leveraged COLI plans, phasing out “otherwise deductible” loan interest by 1999. Without the benefit of the policy loan interest deductions, the COLI VIII plans would not be attractive to corporations.

(2) Expectation of Pre-tax Profitability

Finally, Camelot could not have had a reasonable expectation that its COLI VIII policies would be profitable absent the policy loan interest deductions. As discussed in excruciating, if not nauseating, detail in section III.A.3.a., swpra, prior to purchase of its COLI VIII plan, Camelot had only received sales and issue illustrations showing 20- and 40-year projections, none of which illustrated positive cash flow in the aggregate or for any one year without the policy loan interest deductions. Additionally, as Camelot understood the policy was designed for maximum leveraging by achieving zero net equity by the end of each policy year, it knew that the only benefit of the policy’s cash value was to take out loans that could generate interest deductions, and use the inside build-up as an offset against interest owed. See supra, section III.A.3.a.(2)(a). Moreover, as already discussed, the COLI VIII policies were designed to be mortality neutral. See supra, section III.A.3.a.(2)(b). Therefore, even though Camelot received an unexpected mortality windfall in the early policy years, it cannot credibly argue that it expected to gain from mortality over the life of the plan at the time it purchased the policies.

For the foregoing reasons, the Court concludes that, regardless of how Camelot intended to use the cash flows generated by its COLI VIII plan, the COLI transaction itself had “no purpose beyond creating tax deductions,” and was entered into without an expectation of pre-tax economic profit. ACM, 157 F.3d at 253 (citing Wex-ler, 31 F.3d at 123-24).

c. Cases Cited By Camelot Do Not Support COLI VIII Plan Having Economic Substance

Camelot argues the majority of cases involving the deductibility of life insurance policy loan interest support a conclusion that Camelot’s COLI VIII plan had economic substance, citing the following cases in support of its position: Shirar v. Comm’r, 916 F.2d 1414 (9th Cir.1990); Woodson-Tenent Lab., Inc. v. United States, 454 F.2d 637 (6th Cir.1972); Campbell v. Cen-Tex, Inc., 377 F.2d 688 (5th Cir.1967); Golsen v. United States, 80-2 USTC (CCH) ¶ 9,741, 1980 WL 4732 (Ct.Cl.1980) (“Golsen II”); Coors v. United States, 215 Ct.Cl. 840, 572 F.2d 826 (1978); and Priester Mach. Co. v. United States, 296 F.Supp. 604 (W.D.Tenn.1969). The Court notes, first, that the economic substance determination requires a court to do a probing analysis of the particular transaction at issue. The only case decided thus far addressing the deductibility of interest under a set of facts involving a broad-based leveraged COLI plan is Winn-Dixie, supra.

Woodson-Tenent, Cen-Tex, and Priester each involved the deductibility of interest on policy loans with respect to “key man” life insurance policies. See Woodson-Tenent, 454 F.2d at 639-40; Cen-Tex, 377 F.2d at 689; Priester, 296 F.Supp. at 605-06. In each case, on the first day of the first policy year, the policyholder paid the first year premium and the next four annual premiums discounted at 3%. See id. In a simultaneous netting transaction, the policyholder took out a policy loan in an amount approximating the policy value deemed created by the payment of the first five premiums. See id. The policy loan had a fixed interest rate of 4%. See id. In each case, the respective courts found the interest deductions on these policy loans to not be shams in fact or in substance. See Woodson-Tenent, 454 F.2d at 638-39; Cen-Tex, 377 F.2d at 692-94; Priester, 296 F.Supp. at 608-10.

Beyond the stated similarities, these cases are materially distinguishable from this case and, therefore, of little utility in evaluating the economic substance of Camelot’s COLI VIII transaction. Unlike the Camelot COLI VIII plan, each of these cases involved policies purchased for the legitimate business purpose of providing the company with financial insurance in the event that key company officials died. See Woodson-Tenent, 454 F.2d at 638 (purchase of keyman insurance on lives of two principal officers of company was for legitimate business purpose); Priester, 296 F.Supp. at 605 (“purpose of obtaining such insurance was the usual one for obtaining ‘key man’ insurances, i.e., to provide a cushion in the event of the loss of Mr. Priester’s services”); Cen-Tex, 377 F.2d at 693 (finding a bona fide business purpose to insurance policies purchased for the “objective of having insurance upon its key executives and stockholders” as well as meeting obligations under deferred compensation plan and stock option agreements).

Moreover, in Cen-Tex, Priester, and presumably Woodson-Tenent, 83 even with the policy loans, the policies would have built-up significant net equity, that is unencumbered cash value, in addition to net death benefits. See Cen-Tex, 377 F.2d at 693; Priester, 296 F.Supp. at 606. Shirar, which involved personal rather than corporate-owned life insurance policies, is distinguishable on this basis as well. See 916 F.2d at 1418. In Camelot’s case, by contrast, its COLI VIII plan was designed not to accumulate any significant unencumbered cash value over any one year or the life of the policy.

Coors is inapposite. In Coors, the four adjustable whole life policies at issue were purchased on the life of a family breadwinner with five children for the primary purpose of protecting the family financially against his untimely death. The court found that such protection was “substantial and of economic significance” to his family. Coors, 572 F.2d at 833. By contrast, Camelot’s COLI VIII policies were not purchased to insure against the risk of any employees’ untimely death, but rather were intended to generate predictable cash flows for each and every policy year. Cf. id. at 834 (“There is no evidence in the record that plaintiffs purchased these policies for investment purposes.”). In addition, the policyholder paid cash premiums for the first several years of its policies before taking out policy loans to pay premiums in later years after falling on tough economic times. See id. at 828-29. Un-like the policyholder in Coors, Camelot never once paid an annual premium in cash, instead paying only approximately 10 to 12 percent in cash. Also, in Coors, the cash payments exceeded loan interest owed in each year. Id. at 835. In contrast in Camelot’s COLI VIII plan, cash payments constituted only approximately 40% of loan interest owed. G973. Finally, in Coors and unlike this case, there was no evidence in the record that the deductibility of the policy loan interest played any role in the policyholder’s decision to take out policy loans. Id. at 829.

Most importantly, none of the cases cited by Camelot in support of its policy loan interest deductions involves the sort of complicated, contrived plan designed to generate predictable positive cash flows based on precise calibrations of various plan elements including high premium levels, high interest rates, large numbers of insureds, and pre-orchestrated combinations of loans, loading dividends, and policy withdrawals. Cf. Coors, 572 F.2d at 838-39 (noting the absence of “other significant factors” which could serve “to contaminate the loan transaction and resulting claimed interest deduction,” such as: “... prepayment and/or discounting of premium payments, setting premium payments at an artificially high level in order to facilitate immediate and significant borrowing, ... and contrived, ingenious, and complex loan arrangements whose only ultimate and/or realistic purpose was to secure intended tax deduction benefits.”); cf. also id. at 834 (“There is no basis, on this record, for viewing the loan transactions between [the policyholder and the insurance company] as contrived, ingenious or complex.”). Because the facts of the cases cited by Camelot in support of its position are materially different from Camelot’s broad-based leveraged COLI VIII plan, the Court concludes they are of little utility, and hardly dispositive, in assessing the substance of the transaction at issue in this case. 84

d. Conclusion

On the basis of the cash flow and earnings illustrations which it actually received before and contemporaneously with its purchase of the COLI policies, and which reflected the intended operation of the plan, the Court concludes Camelot could not expect, either in any individual year or over the life of its COLI VIII plan, to achieve positive cash flows or earnings absent the tax benefit afforded by the policy loan interest deductions. Additionally, Camelot did not intend to, nor could it reasonably expect to, profit from cash value build-up in its COLI VIII policies, since the plan contemplated the policies, at each policy anniversary, would have zero net equity. Further, Camelot did not intend to, nor could it reasonably expect to, profit from the death benefits it received on the deaths of its covered employees, since the cost-of-insurance charges under its COLI VIII plan were priced and the plan was operated so as to achieve mortality neutral financial performance.

The only financial benefit provided to Camelot by its COLI VIII plan, and the reason the transaction was profitable to all involved except the United States Treasury, was the tax benefit attributable to policy loan interest deductions. Moreover, Camelot tailored the elements of its COLI VIII plan to maximize the interest deductions to the extent it expected to have taxable income to offset the deductions. The Court concludes Camelot’s COLI transaction lacked objective economic substance and a subjective business purpose apart from the interest deductions. 85 As a consequence, all the interest deductions generated by the plan will be disallowed as the product of a sham transaction.

B. Generic or “Abusive” Tax Shelter

The IRS seeks to persuade the Court to find Camelot’s COLI VIII transaction to be a “generic tax shelter” devoid of economic substance under a test first articulated by the Tax Court in Rose v. Comm’r, 88 T.C. 386, 412, 1987 WL 49274 (1987), aff'd 868 F.2d 851 (6th Cir.1989). In Rose, the Tax Court sought to develop, as an alternative to the traditional two-prong subjective/objective sham transaction test, a unified test, emphasizing objective factors, that could be used to identify tax shelters. The Tax Court outlined five characteristics, some or all of which, generic tax shelters may possess:

(1) Tax benefits were the focus of promotional materials; (2) the investors accepted the terms of purchase without price negotiation; (3) the assets in question consist of packages of purported rights, difficult to value in the abstract and substantially overvalued in relation to tangible property included as part of the package; (4) the tangible assets were acquired or created at a relatively small cost shortly prior to the transactions in question; and (5) the bulk of the consideration was deferred by promissory notes, nonrecourse in form or in substance.

Id. at 412-13. After concluding the transaction at issue was a “generic tax shelter,” the Tax Court went on to determine whether the transaction had economic substance by analyzing: (1) the dealings between petitioners and the promoters; (2) the relationship between the sales price and the fair market value; (3) the structure of the financing; and (4) perceived congressional intent. See id. at 415-22.

The Third Circuit Court of Appeals has never spoken on the Tax Court’s “generic tax shelter” test, and has instead employed the traditional two-prong objective/subjective sham transaction analysis. See ACM, 157 F.3d at 247-48. Moreover, no Circuit Court of Appeals has explicitly adopted the generic tax shelter test. See Nickeson v. Comm’r, 962 F.2d 973, 976 (10th Cir.1992) (while not explicitly adopting the generic test, finding no error in the Tax Court having done so because “the Rose test is simply an application of factors that courts typically use to determine whether investments in activities asserted to be profit motivated were actually entered into for tax benefits”); Hunt v. Comm’r, 938 F.2d 466, 471 n. 5 (4th Cir.1991) (finding “no need to adopt the ‘generic tax shelter’ test, since [the Fourth Circuit’s] two-pronged [business purpose/economic substance] test accomplishes the same purpose”); Smith v. Comm’r, 937 F.2d 1089, 1094 (6th Cir.1991) (noting that the Sixth Circuit did not adopt the “generic tax shelter test” in its affirmance of the Tax Court’s decision in Rose); Karr v. Comm’r, 924 F.2d 1018, 1023 (11th Cir.1991) (finding that the Tax Court did not err in applying the generic tax shelter test because it “correctly focused its analysis upon an examination of the substance of the transactions to determine whether those transactions had any practicable economic effect other than the creation of tax benefits”); Rose, 868 F.2d at 853 (declining to adopt the “generic tax shelter” analysis “because it does not aid [the Court] in the basic inquiry as to whether the transaction has any practicable economic effects other than the creation of income tax losses,” which is “the proper standard”); Collins v. Comm’r, 857 F.2d 1383, 1386 (9th Cir.1988) (“Although the generic tax shelter analysis is not incorrect,” the Court declined to adopt it because “it does not aid courts in [the] basic inquiry”).

There also is considerable doubt as to whether the Tax Court will continue to apply the generic tax shelter doctrine. See Peat Oil & Gas Assocs. v. Comm’r, 100 T.C. 271, 276-77, 1993 WL 95582 (1993) (“We need not here decide whether we will follow the Rose approach in the future in view of its failure to gain acceptance by the Court of Appeals for the Sixth Circuit and others.”). In Peat, eight tax court judges advocated explicitly abandoning the generic tax shelter doctrine. See id. at 279-80 (Swift, J., concurring) (noting that no Court of Appeals has adopted the “unified” generic tax shelter analysis of Rose and stating “I would no longer follow Rose”); id. at 287 (Ruwe, J., concurring) (“[B]ecause the unified approach [of Rose ] has been misunderstood, has been rejected by several Courts of Appeals, and is not particularly helpful, I would discontinue using it.” (footnote and citations omitted)).

In arguing that this Court should find Camelot’s purchase of the COLI VIII policies to be a generic or “abusive” 86 tax shelter, the IRS implies that such a finding would be categorically different from the conclusion that the transaction lacked the substance its form represents. However, several Circuit Courts of Appeals have aptly noted the similarities in the factors that the “generic tax shelter” test examines and those examined under the traditional two-prong inquiry into whether a transaction has economic substance. See, e.g., Nickeson, 962 F.2d at 976; Karr, 924 F.2d at 1023. See also McCrary v. Comm’r, 92 T.C. 827, 845, 1989 WL 35568 (1989) (“Rose simply reformulated [the traditional] two pronged test into a unified approach in certain types of cases.”). Thus, as the Court has already evaluated all the facts and circumstances surrounding the purchase and operation of Camelot’s COLI VIII policies and found them to be lacking in economic substance under the two-prong objective/subjective test that has been adopted by the Third Circuit appellate court, the Court declines the IRS’s invitation to employ the alternative “generic tax shelter” analysis.

C. The I.R.C. § 264(c)(1) Four out of Seven Safe Harbor

I.R.C. § 264(a)(3) 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. 26 U.S.C. § 264(a)(3). I.R.C. § 264(c)(1) provides an exception to I.R.C. § 264(a)(3) 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.” 26 U.S.C. § 264(c)(1) (known as the “four out of seven safe harbor” or “safe harbor”).

Implicit in the four out of seven safe harbor test is the requirement that the annual premiums must remain level for the first seven years of the policy. See 26 U.S.C. § 264(c)(1). Indeed, Congress sought “to prevent avoidance of this provision by taking out a contract with very low premiums for the first 4 years, with the premiums being substantially greater thereafter.” S. REP. NO. 830 (1964), reprinted in 1964 U.S.C.C.A.A.N. at 1751-52. Congress achieved this goal by adding a sentence to the end of § 264(c)(1), which provides “if there is a substantial increase in the premiums on a contract, a new 7-year period described in such paragraph with respect to such contract shall commence on the date the first such increased premium is paid.” Although Congress did not anticipate the opposite problem, where the premium would be very high in years one through three, and then very low in years four through seven, it manifested an intent that premiums be level throughout the first seven years of the policy. A contrary result would allow a policyholder to avoid this provision by taking out loans to pay large premiums in the first three years, and then paying very small cash premiums in the next four years. Such an absurd result should be avoided. Finally, Treasury Regulation 1.264^4(e)(l)(ii) reflects the implicit level premium requirement by stating: “if the stated annual premiums due on a contract vary in amount, borrowing in connection with any premium, the amount of which exceeds the amount of any other premium, on such contract may be considered borrowing to pay premiums for more than one year.” 26 C.F.R. 1.264 — 4(c)(l)(ii). It follows, I.R.C. § 264(c)(1) has an implicit requirement that the premiums remain level during the first seven years of the policy. Camelot has not maintained otherwise.

The Camelot COLI VIII policy required Camelot to pay premiums of $10,000 per insured employee, per year. J189 at 001135. As discussed supra, section III. A.2.a., in the first three policy years, Camelot paid real annual premiums through a combination of real policy loans and real cash transfers. In the fourth through seventh policy years, Camelot purportedly paid the annual premiums through a combination of partial withdrawals, loading dividends, and wire transfers. However, as discussed, supra, section III.A.2.b., the loading dividends were factual shams. Accordingly, there is a substantial shortfall in the payment of annual premiums due in years four through seven:

[[Image here]]

J138; G982. The shortfall illustrated in the above table means Camelot failed to pay about 85% of the annual premium due in years four through seven. Because of this substantial decrease in the amount of premiums actually paid in years four through seven, Camelot has failed to qualify for the § 264(c)(1) four out of seven safe harbor provision. It follows, Camelot’s interest deductions attributable to its policy loans made in the first three policy years must be disallowed pursuant to § 264(a)(3). 87

Camelot relies on Shirar v. Comm’r, 916 F.2d 1414 (9th Cir.1990), in which a life insurance policyholder obtained policy loans to pay premiums in the first three policy years, and then canceled the policy. The Court held the borrowing pattern satisfied the § 264(c)(1) four out of seven safe harbor because the policyholder did not intend to pay the next four premiums by policy loans. See id. at 1419. While it is not clear from the opinion how the policyholder was going to pay the next four years of premium payments, the Court noted the policy was canceled because it was no longer needed to meet potential estimated estate tax obligations because of an intervening change in the tax law. See id. at 1416, 1419. Thus, Shirar fails to support Camelot’s position.

D. Penalties

I.R.C. §§ 6662(a), 6662(b)(2) impose an accuracy related penalty on any portion of underpayment of income tax attributable to a “substantial understatement of income tax.” 26 U.S.C. §§ 6662(a), 6662(b)(2). For corporations, a “substantial understatement” is an understatement of 10% of the tax or $10,000, whichever is greater. 26 C.F.R. § 1.6662-4(b). It is undisputed that Camelot’s understatement of tax on its 1991, 1992, and 1993 tax returns exceeded the greater of 10% of the tax or $10,000. Camelot relies on two statutory exceptions to the “substantial understatement” penalty provision: the substantial authority exception under I.R.C. § 6662(d)(2)(B)(i), and the reasonable cause and good faith exception under I.R.C. § 6664(c)(1). In addition, Camelot relies on a common law exception which it argues disallows penalties if the case involves an issue of first impression. See Mitchell v. Comm’r, 2000 WL 428644, T.C. Memo 2000-145 (April 21, 2000). These three exceptions will be addressed in turn.

1. Substantial Authority Exception Under I.R.C. § 6662(d)(2)(B)(i)

I.R.C. § 6662(d)(2)(B)(i) provides the taxpayer shall not suffer a penalty for any portion of the understatement attributable to “the tax treatment of any item by the taxpayer if there is or was substantial authority for such treatment.” 26 U.S.C. § 6662(d)(2)(B)(i). Treasury Regulation § 1.6662-^l(d)(2) sets forth the standard for determining whether the taxpayer has substantial authority for its position:

The substantial authority standard is an objective standard involving an analysis of the law and application of the law to relevant facts. The substantial authority standard is less stringent than the more likely than not standard (the standard that is met when there is a greater than 50-percent likelihood of the position being upheld), but more stringent than the reasonable basis standard as defined in § 1.6662-3(b)(3).

26 C.F.R. § 1.6662-4(d)(2). For there to be substantial authority, the “weight of the authorities supporting the treatment is substantial in relation to the weight of authorities supporting contrary treatment.” 26 C.F.R. § 1.6662 — 4(d)(3)(i); see also Custom Chrome, Inc. v. Comm’r, 217 F.3d 1117, 1127-28 (9th Cir.2000). The taxpayer bears the burden of proving there is substantial authority. Hitchins v. Comm’r, 103 T.C. 711, 719, 1994 WL 711926 (1994)(taxpayer bears burden of proof on penalties). 88

Substantial authority must come from one or more of an enumerated list of authorities:

[applicable provisions of the Internal Revenue Code and other statutory provisions; proposed, temporary and final regulations construing such statutes; revenue rulings and revenue procedures; tax treaties and regulations thereunder, and Treasury Department and other official explanations of such treaties; court cases; congressional intent as reflected in committee reports, joint explanatory statements of managers included in conference committee reports, and floor statements made prior to enactment by one of a bill’s managers; General Explanations of tax legislation prepared by the Joint Committee on Taxation (the Blue Book); private letter rulings and technical advice memoranda issued after October 31, 1976; actions on decisions and general counsel memoranda issued after March 12, 1981 (as well as general counsel memoranda published in pre-1955 volumes of the Cumulative Bulletin); Internal Revenue Service information or press releases; and notices, announcements and other administrative pronouncements published by the Service in the Internal Revenue Bulletin.

26 C.F.R. § 1.6662-4(d)(3)(iii). Specifically excluded from the substantial authority inquiry are: “[cjonclusions reached in treatises, legal periodicals, legal opinions or opinions rendered by tax professionals.” Id. In addition, “a case or revenue ruling having some facts in common with the tax treatment at issue is not particularly relevant if the authority is materially distinguishable on its facts, or is otherwise inapplicable to the tax treatment at issue.” 26 C.F.R. § 1.6662-4(d)(3)(ii); see also Antonides v. Comm’r, 893 F.2d 656, 660 (4th Cir.1990).

For substantial authority, Camelot relies on the following cases: Woodson-Tenent Lab., Inc. v. United States, 454 F.2d 637 (6th Cir.1972); Campbell v. Cen-Tex, Inc., 377 F.2d 688 (5th Cir.1967); Shirar v. Comm’r, 916 F.2d 1414 (9th Cir.1990); Priester Mach. Co. v. United States, 296 F.Supp. 604 (W.D.Tenn.1969); Golsen v. United States, 80-2 USTC (CCH) ¶ 9,741, 1980 WL 4732 (Ct.Cl.1980) (“Golsen II”)-, Coors v. United States, 215 Ct.Cl. 840, 572 F.2d 826 (1978). 89 However, none of these cases provides substantial authority because each is materially distinguishable. 90

As discussed in section III.A.3.C., supra, these cases do not provide substantial authority to support Camelot’s contention the COLI VIII plan has economic substance. In addition, these cases do not support Camelot’s position that the loading dividends are real because none of these cases involved transactions where the substantial majority of the premiums were “paid” by dividends. See id. Further, none of these cases provide substantial authority that Camelot’s COLI VIII plan satisfies the four out of seven safe harbor of I.R.C. § 264(c)(1). Woodson-Tenent, Cenr-Tex, Priester, Golsen II, and Coors all involved life insurance policies purchased before the effective date of the amendment to I.R.C. § 264 to limit borrowing to those policies which satisfy the four out of seven safe harbor. In addition, as discussed in section III.C., supra, Shirar provides no support for Camelot’s COLI VIII plan satisfying the four out of seven safe harbor of I.R.C. § 264(c)(1). Lacking substantial authority to support its claim for interest deductions on its policy loans, Camelot cannot avail itself of this exception to the imposition of accuracy related penalties.

2. Reasonable Cause and Good Faith Exception Under I.R.C. § 6664(c)(1)

I.R.C. § 6664(c)(1) provides an exception to imposition of an accuracy related penalty when “it is shown that there was a reasonable cause for such position and that the taxpayer acted in good faith with respect to such position.” 26 U.S.C. § 6664(c)(1). As with the prior exception, the nature of this inquiry varies depending on whether the transaction is considered to be a “tax shelter item.” A “tax shelter item” is an item which “is directly or indirectly attributable to the principal purpose of a tax shelter to avoid or evade Federal income tax.” 26 C.F.R. § 1.6662-4(g)(3). Again, the Court need not decide whether Camelot’s policy loan interest deductions are tax shelter items because the result of the inquiry is the same.

If Camelot’s policy loan interest deductions are tax shelter items, in order for this exception to apply, the taxpayer must, at a minimum, satisfy an “authority requirement” and a “belief requirement.” 26 C.F.R. § 1.6664 — 4(e)(2)(i). “The authority requirement is satisfied only if there is substantial authority (within the meaning of § 1.6662-4(d)) for the tax treatment of the item.” 26 C.F.R. § 1.6664-4(e)(2)(i)(A). As discussed in the section III.D.l., supra, Camelot has no substantial authority to justify its COLI VIII plan policy loan interest deductions. Because Camelot lacks substantial authority it cannot satisfy the requirements of the reasonable belief and good faith exception for tax shelter items under I.R.C. § 6664(c)(1).

If Camelot’s policy loan interest deductions are not tax shelter items:

[t]he determination of whether a taxpayer acted with reasonable cause and in good faith is made on a case-by-case basis, taking into account all pertinent facts and circumstances.... Generally, the most important factor is the extent of the taxpayer’s effort to assess the taxpayer’s proper tax liability. Circumstances that may indicate reasonable cause and good faith include an honest misunderstanding of fact or law that is reasonable in light of all of the facts and circumstances, including the experience, knowledge, and education of the taxpayer. An isolated computational or transcriptional error generally is not inconsistent with reasonable cause and good faith. Reliance on an information return or on the advice of a professional tax advisor or an appraiser does not necessarily demonstrate reasonable cause and good faith. Similarly, reasonable cause and good faith is not necessarily indicated by reliance on facts that, unknown to the taxpayer, are incorrect. Reliance on an information return, professional advice, or other facts, however, constitutes reasonable cause and good faith if, under all the circumstances, such reliance was reasonable and the taxpayer acted in good faith.

26 C.F.R. § 1.6664 — 4(b)(1) (emphasis added).

In addition, there are several specific rules for when it is appropriate for a taxpayer to rely in good faith on the opinion or advice of others:

All facts and circumstances must be taken into account in determining whether a taxpayer has reasonably relied in good faith on advice (including the opinion of a professional tax advisor) as to the treatment of the taxpayer (or any entity, plan, or arrangement) under Federal tax law. However, in no event will a taxpayer be considered to have reasonably relied in good faith on advice unless the requirements of this paragraph (c)(1) are satisfied. The fact that these requirements are satisfied will not necessarily establish that the taxpayer reasonably relied on the advice (including the opinion of a professional tax advisor) in good faith. For example, reliance may not be reasonable or in good faith if the taxpayer knew, or should have known, that the advisor lacked knowledge in the relevant aspects of Federal tax law.

(i) All facts and circumstances considered. The advice must be based upon all pertinent facts and circumstances and the law as it relates to those facts and circumstances. For example, the advice must take into account the taxpayer’s purposes (and the relative weight of such purposes) for entering into a transaction and for structuring a transaction in a particular manner. In addition, the requirements of this paragraph (c)(1) are not satisfied if the taxpayer fails to disclose a fact that it knows, or should know, to be relevant to the proper tax treatment of an item.

(ii) No unreasonable assumptions. The advice must not be based on unreasonable factual or legal assumptions (including assumptions as to future events) and must not unreasonably rely on the representations, statements, findings, or agreements of the taxpayer or any other person. For example, the advice must not be based upon a representation or assumption which the taxpayer knows, or has reason to know, is unlikely to be true, such as an inaccurate representation or assumption as to the taxpayer’s purposes for entering into a transaction or for structuring a transaction in a particular manner.

26 C.F.R. § 1.6664-4(c)(l).

Camelot asserts it performed an extremely diligent and thorough three year investigation into the purchase of broad-based leveraged COLI and reasonably and in good faith concluded it had the right to claim interest deductions for the policy loans. Although Camelot conducted a thorough investigation into the COLI VIII plan, the evidence shows Camelot knew, or reasonably should have known, the COLI VIII plan had no practical economic effect other than the creation of tax deductions. Camelot knew that taking policy loan interest deductions, under the COLI VIII plan, would be, at best, a risky move in light of the Internal Revenue Code and the sham transaction doctrine.

As previously set forth in section II.C., supra, Newport communicated to Camelot that the primary function of the COLI VIII policy was to fuel policy loan interest deductions. J7; J15 at Bates 000461; Campisi, Tr. 1405-12. Also, all of the sales illustrations Campisi sent to Camelot showed that the policy would have a negative cash flow absent the policy loan interest deductions. See, supra, section III. A.3; see also J188, J17, J33, J35/G403. In addition, as previously rehearsed in section III.A.3.b.(l), supra, the principal decision makers at Camelot understood that the overarching consideration behind the COLI VIII plan was to create interest deductions. Rogers, Tr. 406-09, 424, 427, 450-52, 487-89; Thorn, Tr. 668-70, 696, 699. This is further evidenced by Camelot’s efforts to avoid the potential impacts of adverse tax legislation eliminating the policy loan interest deduction. Rogers, Tr. 244, 484-85, 494; Tr. J22; J23; J26.

Camelot asserts it reasonably relied in good faith on an analysis by its internal tax director, Thorn, who determined the COLI VIII policy complied with the $50,000 per policy borrowing limit and the four out of seven safe harbor in I.R.C. § 264. Thorn, Tr. 669-73; 677, 669-73. However, Thorn did not conduct an investigation into whether the policies were shams lacking in economic substance. She did not consider whether the loading dividends in years four through seven were real. She also did not understand that there was a crediting rate on the amount of policy value equal to the loans. Thorn, Tr. 704-05. Given Thorn’s rather limited investigation and ignorance of how the policy worked, and the abundance of knowledge concerning how the COLI policy simply fueled interest deductions, Camelot was not entitled to rely in good faith on Thorn’s legal conclusions.

In addition, Camelot urges it relied in good faith on Ernst & Young’s, Camelot’s outside auditor, approving Thorn’s analysis and Ernst & Young’s reviewing and signing Camelot’s tax returns claiming the interest deductions. Thorn, Tr. 673-74. Rogers, Tr. 192-94, 217-18. But, Camelot ignores Ernst & Young warned Camelot the COLI VIII plan was a tax-driven investment subject to attack by the IRS:

It should be recognized that the proposed COLI program is designed to take advantage of a tax arbitrage opportunity existing under present tax law. Even though the program may qualify as an insurance contract, it is, in reality a tax-advantaged investment program, the overall economics of which generates a substantial internal rate of return and cash flow from the deductibility of the interest paid on borrowed CSV [cash surrender value]. As such, there is a degree of tax risk that the IRS will challenge the program and deductibility of the interest paid from a variety of standpoints.

GX 374, Bates 000081; see also Rogers, Tr. 498-99. Thus, despite the fact Ernst & Young signed off on Thorn’s analysis and Camelot’s tax returns, it made Camelot well aware of the potential risks of an IRS attack. As such, Camelot cannot rely in good faith on advice by Ernst & Young. 91

Finally, there is substantial evidence that Camelot knew that the COLI policy was fraught with tax risk and chose to enter the transaction anyway. This fact is nowhere more clearly shown than in Jack Rogers’ memo to the file on the day Camelot decided to purchase the COLI VIII product. J25. The Court again stresses, in that memo, he wrote:

Overall, we have satisfactorily resolved the legal issues involved with this plan. The decision to go forward and institute a COLI plan boil down to the risks involved. Those risks specifically include:

1) A retroactive tax law change

2) Camelot’s failure to generate taxable income over several years in a row

3) IRS attack

J25. The only risks identified in this memo are those related to whether Camelot would continue to be able to take the policy loan interest deductions. Rogers, Tr. 245-47, 501, 507. Camelot, a sophisticated corporate business, conducted a three year due diligence investigation and concluded its COLI VIII plan contained a risk the policy loan interest deductions would be disallowed by the IRS. Having knowingly taken such a risk, it cannot be said Camelot acted with reasonable cause and in good faith in taking the policy loan interest deductions. It follows, Camelot is not entitled to the reasonable cause and good faith exception to the imposition of penalties.

3. Common Law Issue of First Impression Exception

Camelot relies on a purported common law exception to accuracy related penalties which prohibits imposition of penalties “in a case of first impression involving the unclear application of an amendment to the Internal Revenue Code.” Mitchell, 2000 WL 428644, T.C. Memo 2000-145 (citing Bunney v. Comm’r, 114 T.C. No. 17, 2000 WL 380121 (2000); Lemishow v. Comm’r, 110 T.C. 110, 114, 1998 WL 130156 (1998); Hitchins v. Comm’r, 103 T.C. at 720; Everson v. United States, 108 F.3d 234, 237-38 (9th Cir.1997)). The Court finds this exception inapplicable for two reasons.

First, Mitchell was the first and only case to apply this exception to accuracy related penalties for substantial understatement. See Mitchell, 2000 WL 428644, T.C. Memo 2000-145. In contrast, all prior cases, including those cited in Mitchell, applied this exception only to negligence based penalties. See Bunney, 114 T.C. No. 17, 2000 WL 380121; Lemishow, 110 T.C. at 114; Hitchins, 103 T.C. at 719-20; Everson, 108 F.3d at 237-38. Extending this exception to accuracy related penalties for substantial understatement is of dubious utility because it would undermine the purpose of accuracy related penalties, which is to penalize “taxpayers who take questionable ... positions not amounting to fraud or negligence on their returns in the hope that they will not be audited.” Staff of Joint Comm, on Taxation, 97th Cong., General Explanation of the Tax Equity and Fiscal Responsibility Act of 1982, at 215-17 (Comm. Print 1982); see also H.R. REP. NO. 101-247, at 1389-90 (1989), reprinted in 1989 U.S.C.C.A.N. 2859-60 (1989). Accordingly, the Court declines to follow the Tax Court in Mitchell, and apply this exception to accuracy related penalties for substantial understatement.

Second, even if this first impression exception were to apply to accuracy related penalties for substantial understatement, it would not be applicable in this case. To satisfy this exception the issue must be “one not previously considered by the Court and the statutory language was not entirely clear.” Hitchins, 103 T.C. at 719 (emphasis added); see also Bunney, 114 T.C. No. 17, 2000 WL 380121; Lemishow, 110 T.C. at 114. Camelot points to this Court’s opinion on the withdrawal of reference motion, where this Court held the case involves “two unsettled tax issues of first impression which require substantial and material consideration of federal tax law.” CM Holdings I, 221 B.R. at 723.

The first issue is whether the policy loans constitute genuine indebtedness under I.R.C. § 163 or whether they are sham transactions, in fact or in substance. See id. at 722. Although this is an issue of first impression as it relates to the COLI VIII policy loans, it does not involve interpretation of unclear statutory language. See id. (citing United States v. Wexler, 31 F.3d at 122); see also, supra, section III.A. Instead, it involves applying novel facts to the judicially created sham transaction doctrine. See id. It follows, this first impression exception should not apply to the sham transaction issue.

The second issue is whether the policy loans qualify for the four out of seven safe harbor in I.R.C. § 264(c)(1). In the withdrawal of reference opinion, this Court held this issue was an issue of pure statutory interpretation “of the operative language ‘annual premiums due’ ” and whether it included premiums or premiums less dividends and partial withdrawals. CM Holdings I, 221 B.R. at 723. However, at trial and in the post trial submissions, the parties now dispute the factual issue of whether the annual premiums paid by loading dividends are real. That decision then dictates whether Camelot has satisfied the four out of seven safe harbor. See, supra, section III.C. This issue does not involve any legal interpretation of the meaning of the statute. As stated previously, the Court misconstrued the nature of the inquiry in its earlier opinion or the IRS abandoned its legal line of attack in favor of a factual attack. In either event, this issue does not involve interpretation of unclear statutory language. It follows, it would be inappropriate to apply the first impression exception in this case.

Since Camelot cannot avail itself of any of these three exceptions, imposition of accuracy related penalties for substantial understatement, pursuant to I.R.C. §§ 6662(a), 6662(b)(2), is appropriate.

Before concluding comment must be made on the quality of the litigants’ representation. Counsel for both the IRS and Camelot provided their respective clients superior trial and advocacy skills. During this high tech trial, the attorneys exhibited complete mastery of the voluminous exhibits, including the ability to quickly locate virtually any document. They cooperated with each other to the fullest extent possible, so long as it did not materially harm the interest of their respective clients. Moreover, their civility in the courtroom was beyond reproach. Not only was the trial high tech, but each litigant provided the Court with a CD-ROM which contained their lengthy, but excellent, post-trial submissions with electronic links to the transcript, exhibits and cases. This electronic submission enabled the Court to more efficiently review the post-trial papers and the record. One may comfortably predict that CD-ROM’s with electronic links will at some point in the future become the norm in post-trial submissions of lengthy, complex bench trials.

IV. CONCLUSION

Camelot, in going forward with its COLI VIII plan, crossed the line separating a life insurance contract where policy loan interest is deductible and a tax advantaged investment program where it is not. The Court has rejected the government’s position that policy loans, partial withdrawals, and payment of annual premiums during the first three policy years were factual shams. However, purported payment of premiums during policy years four through seven have been held to be factual shams, because they were primarily funded by loading dividends which were the product of circular accounting treatment. Because premium payments for policy years four through seven were factual shams, there was no level premium during the first seven years of the Camelot COLI VIII plan leading the Court to hold Camelot failed to qualify for the § 264(c)(1) four out of seven safe harbor provision, thereby making the interest payments non-deductible.

The Court also found that absent the interest deduction, Camelot could not reasonably expect to realize a positive cash flow during any year of the COLI VIII plan. Further, the COLI VIII plan was designed so that Camelot could not expect to benefit from the inside cash value buildup or profit from the death benefits on covered employees. The only financial benefit flowing to Camelot from its COLI VIII plan was the tax benefit attributable to policy loan interest deductions. For these reasons, the Court held Camelot’s COLI VIII plan lacked objective economic substance and a subjective business purpose other than the tax benefits flowing from the interest deductions. In short, Camelot’s COLI VIII plan was a sham in substance which served as an additional basis for disallowing the interest deductions.

In sum, the Court has held the Camelot COLI VIII plan loading dividends primarily used to fund premiums for policy years four through seven are shams in fact. Further, because of the shams in fact which occurred during policy years four through seven, the Court held Camelot failed to meet the § 264 four out of seven safe harbor test. The Court also held the Camelot COLI VIII plan was a sham in substance. For these reasons, the policy loan interest deductions taken by Camelot under its COLI VIII plan have been disallowed. 92 Finally, the Court held imposition of accuracy related penalties for substantial understatement is warranted.

APPENDIX A

APPENDIX B

APPENDIX C

[[Image here]]

[[Image here]]

[[Image here]]

1 .According to the introduction of Petitioner's Proposed Findings of Fact and Conclusions of Law, the IRS assets there is tax due because of the disallowed deductions in the amount of $5,290,554 and also asserts it assessed accuracy related penalties for substantial understatement of income tax, pursuant to 26 U.S.C. § 6662, in the amount of $950,900.40. The Court was unable to find these particular numbers in the voluminous record.

The outcome of this case potentially has far greater consequences than the approximately $6 million allegedly owed by Camelot. In addition to MBL, seven other life insurance companies have sold similar COLI plans to many Fortune 500 companies and mid-sized corporations. George Imwalle ("Imwalle”), Transcript ("Tr.”) 977-80, 3288. At the time of trial, 85 taxpayers with similar plans had been identified by the IRS, 50 investigations were underway, and many more were expected to be added to the list. Imwalle, Tr. 3317, 3287. As a result of these investigations, there is an aggregate potential tax liability of $6 billion. Imwalle, Tr. 977.

2 . The IRS’s exhibits are referred to as G_; the respondent’s, Camelot's, exhibits will be referred to as R_; and the joint exhibits will be referred to as J_When referring to an expert report, the expert’s name will precede the exhibit designation.

3 . In those few instances when depositions were not taken by video, witness demeanor could not be evaluated.

4 . At the request of the Court, the litigants supplied glossaries of terms. Because the definitions are largely undisputed, the Court draws upon the IRS’s glossary, which was prepared by Dan M. McGill, Frederick H. Ecker Emeritus Professor of Life Insurance at the Wharton School of the University of Pennsylvania. This glossary will be used without attribution.

5 . MCGILL’S LIFE INSURANCE was referenced at trial, and a copy provided to the Court, so that the Court could obtain a better general understanding of life insurance. As appears from the text, the Court has heavily drawn upon it in this non-controversial background discussion. However, because it is not in evidence, it will not be relied upon in the more substantive portion of this opinion, unless actually referenced in the transcript.

6 . The disintermediation crisis also led the creation of the Model Bill on Policy Loan Interest Rates by the National Association of Insurance Commissioners ("NAIC”). The Model Bill is discussed in greater detail in section III.A.2.a.(3), infra.

7 . In response to a question as to what an actuary is and does, Eisenberg responded:

An actuary is trained to analyze risks, the types of risks you find particularly with insurance, such as the risks of dying too early, living too long, becoming sick, disabled, having catastrophes, things like that, and how to measure and quantify those risks and put a price on those risks.

Eisenberg, Tr. 2153. Within the life insurance field, one of the primary functions of an actuary is to develop an insurance product. In order to do that, an actuary has to understand economics, math, and statistics and have a working knowledge of contract law and accounting. In addition, an actuary must prepare an actuarial memorandum which becomes part of the filing on seeking state approvals. The actuarial memorandum demonstrates how the policy complies with various regulatory laws for nonforfeiture values and minimum reserves. In addition, it discusses how the product operates with much more detail than found in the policy itself, including what the guarantees in the contract are and how the policy value evolves. Eisenberg, Tr. 2187.

8 . In 1984, Congress enacted I.R.C. § 7702 to define a life insurance contract for purposes of federal tax law as a contract which: (1) is a life insurance contract under applicable state law and (2) satisfies either a statutory cash value accumulation test or meets the guideline premium requirements of § 7702(c) and falls within the cash value corridor test of § 7702(d). 26 U.S.C. § 7702. One witness described § 7702 as governing the relationship between the death benefit and the policy value. DesRochers, Tr. 1877. Another witness described § 7702 as defining what a life insurance policy had to do to be a life insurance policy under the I.R.C. so as to get the tax-free death benefit and tax-deferred cash build-up. Eisenberg, Tr. 2169. Whether Camelot’s COLI VIII policies qualify as life insurance contracts under I.R.C. § 7702 is not an issue in this case.

9 .After she prepared her expert reports, Kristie Sayre became unable to testify at trial. By stipulation of the parties, her husband and business partner, Ralph Sayre, adopted her reports and testified on her behalf at trial. Ralph Sayre also served as an expert witness on other issues for the IRS and testified at trial. For convenience, this opinion will refer to Ralph Sayre’s testimony on behalf of Kristie Sayre as Kristie Sayre's testimony.

10 . While I.R.C. §§ 7702 and 7702A limited the design freedom of the actuary and occupied considerable time at trial, they play only nonsubstantive roles in the resolution of the litigation.

11 . Newport is a national insurance brokerage firm which also provides a range of administrative services for COLI plans. Campisi, Newport’s president, was primarily responsible for the COLI VIII sale to Camelot. Stip. V 31; Campisi, Tr. 1044; R539.

12 . One policy was subsequently rescinded. Stip. ¶ 61.

13 . The Camelot COLI VIII plan consists of 1430 individual policies on each Camelot employee. For convenience, from this point forward the COLI VIII policies will be referred to in the singular unless the context demands otherwise.

14 .The illustration in Appendix A actually shows a first year annual premium of $14,-310,000. This is because, at the time of the projection, Camelot had purchased policies on 1,431 employees. Later, Camelot amended its COLI VIII plan to include only 1,430 COLI VIII policies. Stip. ¶ 61. The COLI VIII plan covered only those Camelot employees. If an employee left Camelot, the policy on his or her life remained in effect until death (a fact ascertainable by a sweep of Social Security numbers) unless Camelot otherwise chose to terminate the COLI VIII plan. Conversely, employees new to Camelot after the plan was purchased are not part of the plan.

15 . Also, in some policy years, there was projected to be a small administrative fee (column (3)). This administrative fee is relatively unimportant to understanding or analyzing the plan.

16 . Camelot did not pay a premium and did not take a policy loan in policy years eight and nine, as originally planned.

17 . Although these transactions actually occurred on April 30, 1990, they were deemed to have occurred on February 16, 1990, the date of the Prepayment Agreement. Campisi, Tr. 1199-1200.

18 . Camelot actually made payments by check to MBL. On February 16, 1990, Camelot paid approximately $278,000 as part of the Prepayment Agreement. Stip. ¶ 55; J26; J28. On April 30, 1990, Camelot paid MBL an additional approximately $708,000, which was deemed to have been paid as of February 16, 1990. Stip. ¶ 64; J39.

19 . In policy year three, the total amounts due were also projected to be paid, in part, by a relatively small cash dividend of $967,000. The mechanism of the cash dividends is discussed, infra, with regard to policy years four through seven.

20 .The values in this chart are a fair representation of the loading charges and anticipated expenses for males and females of most ages insured by Camelot. Eisenberg, Tr. 2478-2483; J165 at Bates 001418. For a 28 year old male, the expense charges are the same in years one through seven but are 68% in years and eight and thereafter. J189 at Bates 001135.

21 . The loaned crediting rate under the original version of FA85, without endorsement EFA85-3, was Moody’s Baa. J189 at Bates 001079.

22 . Although the record contains evidence of policy loans taken on the first day of the policy year, apart from Eisenberg's testimony describing this innovative feature and stating he had heard there was another policy with first day dividends, there is nothing else in the trial record indicating there had ever been a first day "dividend.”

23 . The fact that a tax claim arises in a federal bankruptcy proceeding does not alter the taxpayer's burden of proof. See Raleigh v. Illinois Dep't of Revenue, - U.S. -, 120 S.Ct. 1951, 147 L.Ed.2d 13 (2000) (where the burden of proof is on taxpayer under substantive law creating tax obligation, the burden of proof on the tax claim in bankruptcy court remains where it is under the substantive law); Resyn Corp. v. United States, 851 F.2d 660, 662-63 (3rd Cir.1988).

24 . Finally, Camelot relies upon an e-mail from Teri Culbertson to Nancy Knapp, R99, a chart purporting to show the IRS's treatment of several other COLI plans offered by different carriers, R199, and the deposition testimony of George Imwalle, the Issue Specialist for IRS’s COLI Program, Tr. 956-999, 3284-3324, to demonstrate the IRS has not disallowed interest deductions related to other similar COLI plans. See Respondent's Reply Brief, D.I. 128 at 28. The IRS has objected to the admission of these items of evidence as being irrelevant because the determination of tax liability for a taxpayer must be based on the facts of a particular case, not with reference to other similarly situated taxpayers. See Petitioner’s Legal Argument and Evidentiary Objections, D.I. 118 at 54-55 (citing Hanover Bank v. Comm'r, 369 U.S. 672, 686, 82 S.Ct. 1080, 8 L.Ed.2d 187 (1962); Weller v. Comm'r, 270 F.2d 294, 298 (3d Cir.1959), cert. denied, 364 U.S. 908, 81 S.Ct. 269, 5 L.Ed.2d 223 (1960); Easter House v. United States, 12 Cl.Ct. 476, 490 (1987); Goodstein v. Comm’r, 267 F.2d 127 (1st Cir.1959); Davis v. Comm’r, 65 T.C. 1014, 1022, 1976 WL 3750 (1976); Teichgraeber v. Comm’r, 64 T.C. 453, 1975 WL 3168 (1975); IBM Corp. v. United States, 170 Ct.Cl. 357, 343 F.2d 914 (1965); Penn-Field Indus., Inc. v. Comm’r, 74 T.C. 720, 722, 1980 WL 4468 (1980)). The Court agrees these items are irrelevant to the determination of whether Camelot is entitled to take interest deductions and they will not be relied upon in the Court's sham transaction decision.

25 . Aff'd, 845 F.2d 746 (7th Cir.1988) and aff'd. sub nom., Enrici v. Comm’r, 813 F.2d 293 (9th Cir.1987); Bramblett v. Comm’r, 810 F.2d 197 (5th Cir.1987); Mahoney v. Comm’r, 808 F.2d 1219 (6th Cir.1987); Wooldridge v. Comm'r, 800 F.2d 266 (11th Cir.1986).

26 . As discussed in the Background, section II.D.4, supra, net equity is equal to the policy value less policy loans and accrued interest. When a policy has a net equity equal to zero, all of the policy value is said to be encumbered and the policy value cannot support taking an additional policy loan.

27 . Although McGill originally testified he had previously seen first year, first day policy loans, he later stated he misspoke and clarified that he had only seen first year loans, but not first year, first day loans. Id. The Court credits McGill's revised testimony but reiterates that McGill also recognized the existence of half premium first year loans.

28 . Camelot observes that, because the Current Credited Loaned Interest Rate cannot drop below 4%, there is a risk of disintermed-iation if the variable loan interest rate falls below 4%. Because Camelot elected the Baa enhanced plus T=l% variable loan interest rate, Moody’s Baa would have to drop below 2.88% for the loan interest rate to drop below 4%. In the past 81 years, Moody's Baa has never dropped below 2.94%. Moody’s Industrial Manual, v. 1 at a73 (1999). Therefore, there is virtually no disintermediation risk as the likelihood of a 4% Credited Loaned Interest Rate is minimal or nonexistent.

29 . The interest rates for the policy are determined based upon Moody's publication in the calendar month two months before the date the interest rates were set, which was two months before the start of the policy year. J189 at Bates 001078. Since Camelot’s policy year started in February, the Moody’s pub lished rales for the prior October were used to calculate the rates for the policy.

30 . Having found the loan interest rates to be excessive, the Court declines to address the IRS’s third argument the loan interest rates violate the NAIC Model Bill on Policy Loan Interest Rates.

31 . The IRS asserts the interest was also paid via loading dividends in these years. For a discussion of why the partial withdrawals were themselves sufficient to pay the balance of the accrued interest, see section III.A.2.C, infra.

32 . Camelol relies on testimony by DesRoch-ers and McGill, who both stated it is acceptable in the life insurance industry to net dividends against the payment of premiums on the first day of a policy year. DesRochers, Tr. 1982-83; McGill, Tr. 3797. However, neither witness stated, nor was there any other evidence to support, that netting is an acceptable norm in the life insurance industry if the source of the dividends is the deemed payment of a premium on the same day.

33 . This allusion to surplus is discussed, infra, section III.A.2.b.(2).

34 . Camelot also points to testimony by Bartlett and McGill that there are no statutory, regulatory, or actuarial limitations on the magnitude of the loading charges. Bartlett, Tr. 3518, 3520; McGill, Tr. 3851. However, the absence of these standards does not negate Bartlett’s and McGill's testimony concerning the accepted industry practice of having the loading charges be reasonably related to anticipated expenses and risks.

35 . Camelot points to KENNETH BLACK, JR. & HAROLD J. SKIPPER, JR., LIFE AND HEALTH INSURANCE 795 (13th ed. 2000) (“BLACK & SKIPPER”) to argue the magnitude of Camelot's expense charges was not excessive. Camelot has not provided an exhibit or transcript cite to this textbook. The Court declines to sift through the voluminous exhibits and almost 5,000 pages of transcript to locate this reference. Camelot did provide a copy of the cited page of BLACK & SKIPPER in its CD-ROM, but that does not make it evidence. Nonetheless, even if this document were in evidence, it would not carry the day because it uses an example of an expense charge which is 29% of the premium and a margin which is 4.5% of the premium, values substantially less than those in COLI VIII.

36 . The IRS also challenges the pattern of the loading charges as departing from industry norms because the loading charges are low in the first three years and then increase dramatically in the fourth through seventh years. In fact, Eisenberg forthrightly answered a question posed by tire Court that he had never seen any other policy with this pattern of loads. Credible evidence was presented that, in most life insurance products, it is normal to have very high loading charges (close to 100%) in the first year to pay broker commissions, followed by a steady decline in later years. Eisenberg, Tr. 2276-78; McGill, Tr. 3630-34. Although there is a wealth of evidence that the magnitude of the loading charges in years four through seven was excessive, there is little evidence to support the assertion that the loading charges should have been large in early years, followed by a decline.

37 . Although it may be problematic, the IRS has not challenged, and the Court does not rule on the reality of the first year, first day dividend. The Court received no evidence to explain the origins of this dividend.

38 . Under this endorsement, the expense charge in the third and later policy years was 100% of the premium. A guaranteed dividend would be credited at the end of each policy year equal to 104% of the premium less the policy fee multiplied by the difference between the large premium expense charges and actual policy expenses. The effect of the proposed endorsement was to guarantee the payment of a dividend in those years in which policy loans could not be used to offset premiums and loan interest. Id.

39 . COLI V was an earlier version of COLI VIII which also used policy loans to pay premiums in years one through three and loading dividends to pay premiums in years four through seven. Eisenberg, Tr. 2183-85; G542 at 3-4.

40 . Contrary to Camelot's objection based upon lack of relevance, the evidence related to representations MBL and IAS made to persons other than Camelot is relevant because it demonstrates how the designers of the MBL COLI VIII policy attempted to guarantee the payment of dividends.

41 . The federal DAC tax is an acceleration of the insurance company’s taxable income from premiums. 26 U.S.C. § 848; see also Van Etten, Tr. 2724-26.

42 . This does not include dividends credited to policy value, such as the first year, first day dividend.

43 . During design of the COLI VIII policy, MBL officers sought to avoid creating a liability for the loading dividends by recharacteriz-ing them as "premium rate adjustments.” G89 at Bates H0040. Van Etten rejected this idea as "not viable” since "it would reopen sec. 264 issues with all of our existing clients with potentially disastrous results.” G89 at Bates H0040; Van Etten, Tr. 2911-2917.

44 . In addition, the IRS urges the Court to consider G177, an exhibit reflecting the concerns of James L. Devine of the New York State Department of Insurance about the factual reality of the loading dividends. See Petitioner’s Legal Argument and Evidentiary Objections, D.I. 118 at 52. At trial, Camelo! objected to the relevance of this document and the Court provisionally sustained the objection, relying on its in limine ruling that the opinions of state regulators are irrelevant to determining whether the transaction is a sham. Tr. 2509-13. The IRS urges the Court to revisit this ruling. Having found the loading dividends to be factual shams absent this evidence, the Court finds it unnecessary to revisit its ruling at trial or its in limine opinion.

45 . I.R.C. § 808 provides, in pertinent part:

(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,

(2) excess interest,

(3) premium adjustments, and

(4) experience-rated refunds.

(d) Definitions. — For purposes of this section—

(1) Excess interest. — The term "excess interest” means any amount in the nature of interest—

(A) paid or credited to a policyholder in his capacity as such, and

(B) in excess of interest determined at the prevailing State assumed rate for such contract.

(2) Premium adjustment. — The term "premium adjustment” means any reduction in the premium under an insurance or annuity contract which (but for the reduction) would have been required to be paid under the contract.

(3) Experience-rated refund. — The term "experience-rated refund” means any refund or credit based on the experience of the contract or group involved.

26 U.S.C. § 808.

46 . The IRS argues the "power of the fiction of the partial withdrawal mechanism ... is demonstrated by the eighth-year force-out transaction.” Petitioner’s Legal Argument and Ev-identiary Objections, D.I. 118 at 25. In the eighth year, Camelot took a partial withdrawal of approximately $30 million, which it simultaneously used to repay an equivalent amount of policy loan interest and principal. Van Etten, Tr. 2846-48; Kristie Sayre, Tr. 4050-53, 4056-61, 4062-64, R314/G738 at 43-44; J138, G791. Although the Court need not and does not rule on the reality of the eighth year partial withdrawal, the Court notes the policy value was sufficient in year eight to support this partial withdrawal. As such, the eighth year force out transaction does not provide evidence the partial withdrawals in years four through seven were factual shams.

47 . As noted, supra, section III.A.2.C., the partial withdrawals in years four through seven were used to pay the accrued loan interest.

48 . The Winn-Dixie court decided, without benefit of analysis, the transaction was not a sham in fact. See id. at 291. In addition, the parties stipulated that Winn-Dixie's pattern of borrowing fell within the four out of seven safe harbor of I.R.C. § 264(c)(1). Further, the Winn-Dixie COLI plan did not involve the use of substantial loading dividends to pay premiums, unlike Camelot's. See generally id.

49 . In its brief, Camelot implies that the Third Circuit appellate court has adopted this position. See D.I. 123, at 58 ("According to the Third Circuit, 'where a transaction has economic substance, it does not become unprofitable merely because it is likely to be unprofitable on a pre-tax basis.’ ” ACM, 157, at 257 n. 49 (quoting [Sadcs ].)). However, the Third Circuit Court of Appeals in ACM was merely quoting the holding in Sacks in its discussion of ACM’s argument that a transaction need not be profitable to be respected for tax purposes. The court went on to note that the transaction in Sacks "had economic substance because it involved a sale and leaseback of equipment used for legitimate business purposes and it resulted in concrete changes in the parties economic positions.” Id. (citations omitted). The court then found Sacks to be "inapposite” to ACM because the transaction in ACM did not serve any non-tax purposes and did not materially alter ACM’s net economic position. Id. The Third Circuit Court of Appeals also went on to uphold the Tax Court’s conclusion that ACM would not have reasonably expected its transaction to yield a pre-tax profit because:

ACM had planned and executed its transactions without regard to their pre-tax economic consequences.... The documents outlining the proposed transactions, while quite detailed in their explication of expected tax consequences, are devoid of such detailed projections as to the expected rate of return on the ... essential [non-tax] components of each proposal.

157 F.3d at 257. Similarly, as will be discussed infra, in this case the evidence overwhelmingly indicates that Camelot, while focusing extensively on the tax consequences of the COLI VIII, paid scant attention to the expected rates of return on the non-tax components of the plan.

50 . Additionally, Sacks is distinguishable from this case in other respects. Sacks involved the allowability of depreciation deductions and investment credits on a sale and leaseback transaction regarding solar energy equipment. The court concluded the transaction was not a sham because it had "genuine economic effects” based on several factors: (1) Sack’s was personally liable on the notes; (2) he paid a fair price; (3) "the tax benefits would have existed for someone, either BFS Solar [the other party to the transaction] or Mr. Sacks, so the transaction shifted them but did not create them from thin air”; (4) installing solar water heaters on homeowners’ roofs was a genuine business; and (5) the consequences of markets shifts in the prices of energy were genuinely shifted to Sacks by the transaction. 69 F.3d at 988. In contrast to Sacks, Camelot’s policy loans were non-recourse obligations. Further there was always policy value that could, through reverse netting, be used to pay off loan principal, as occurred with the $26 million payment of loan principal during the eighth policy year. Further, the interest deductions at issue in this case were solely the creature of Camelot's COLI VIII plan and could not be taken by someone else. Additionally, and as discussed in greater detail infra, Camelot did not have a genuine business purpose apart from the tax deductions that were intended to fuel the positive cash flows of the COLI VIII plan. Finally, as set forth infra, in the highly-structured pre-planned transaction, Camelot did not assume any risks of market shifts as Sacks did. The COLI VIII plan was designed and operated to minimize if not eliminate any risk shifting with regard to mortality. As succinctly stated in a Camelot memorandum, the risks inherent in the COLI VIII plan related, not to any market risks, but instead to the deducti-bility of the policy loan interest deductions, i.e., the possibility of changes to the tax law, Camelot’s inability to generate sufficient taxable income to generate the deductions, and IRS attack. See J25.

51 . Camelot also asserts, under ACM, the expectation of a pre-interest deduction profit is relevant only to analysis of subjective business motivation, to help inform the inquiry into why the taxpayer entered the transaction. See 157 F.3d at 253-54. Camelot asserts the Winn-Dixie court erred by turning pre-inter-est deduction profitability, a subsidiary consideration of the subjective analysis, into the defining inquiry of objective economic substance. Resp. Br. at 69 (citing 113 T.C. at 260-68). Further, Camelot complains the Winn-Dixie court did not explain how pre-interest deduction profitability relates to an objective change in economic position. Id. In so doing, asserts Camelot, the Winn-Dixie court ignored a long-standing maxim of tax law: the "fact that favorable tax consequences were taken into account by [a taxpayer] on entering into the transaction is no reason for disallowing those tax consequences. We cannot ignore the reality that the tax laws affect the shape of nearly every business transaction." Frank Lyon Co. v. United States, 435 U.S. 561, 580, 98 S.Ct. 1291, 55 L.Ed.2d 550 (1978); see also Gregory, 293 U.S. at 469, 55 S.Ct. 266 ("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.”); ACM, 157 F.3d at 248 n. 31 ("[W]e do not intend to suggest that a transaction which has actual, objective economic effects on a taxpayer's non-tax affairs must be disregarded merely because it was motivated by tax considerations.”).

Although the existence of a reasonable expectation of pre-interest deduction profitability is relevant to the subjective prong of the analysis, it is also relevant to the objective analysis. ACM requires the Court to assess, for purposes of an objective economic substance analysis, "whether the transaction has any practical economic effects other than the creation of income tax [deductions].” 157 F.3d at 247 (internal quotation omitted; emphasis added). Implicit in this statement is the requirement that the Court look at the transaction’s profitability apart from the interest deductions. Moreover, as set forth above, the ACM court stated that the economic substance analysis is not a "rigid two-step analysis.” 157 F.3d at 248 (internal quotation omitted). Thus, this mode of analysis is appropriate for assessing the objective economic substance of a transaction.

52 . In Winn-Dixie, the illustrations that the Tax Court relied upon to evaluate the economic substance of the broad-based leveraged COLI plan at issue in that case were the illustrations Winn-Dixie actually received shortly before it purchased the plan. See 113 T.C. at 281.

53 . Camelot also asserts, and the Court agrees, that when analyzing the objective economic aspects of the COLI VIII plan, it is inappropriate for the Court to look to policy illustrations that do not reflect the ultimately purchased plan in terms of, e.g., the number of insureds and the premium level. While the number of insureds decreased from the 1431 shown in some of the illustrations to 1430 employees, it is of no consequence.

54 . The illustrations themselves are dated April 27, 1990. J42.

55 . Because the cash flow and earnings illustrations do not differ materially in a manner that would affect an analysis of the substance of Camelot's COLI VIII plan, and because the focus of Camelot's purchase decision and discussions leading to that decision focused on the cash flows, for simplicity, hereafter the Court will focus on the cash flow illustration.

56 . To the extent Camelot claims its COLI VIII plan had economic substance because Camelot was projected in this illustration to receive $2.177 billion aggregate death benefits, this figure, standing in isolation, is no more an indication of the actual economics of the COLI VIII plan than the even greater $2.5 billion in aggregate loan interest charges that are also shown in these projections. The only true measures of the actual economics of this, or any other, issue illustration are the level and timing of its net cash flows.

57 .Additionally, there is no evidence that Rogers or anyone else at Camelot ever saw these illustrations before this litigation, much less reviewed and relied upon them contemporaneously with the purchase of the COLI VIII plan. Rogers acknowledged his lack of familiarity with these issue illustrations, testifying on both direct and cross-examination that the first time he had ever seen them was shortly before his deposition in February 1999. Rogers, Tr. 298-99, 465-66. In his testimony, Campisi explained that Newport likely did not pass this set of issue illustrations along to Camelot because of a flaw projecting enormous negative after-tax cash flows (approximately $97 million) in plan years 47 through 53 in Scenario 3. Campisi, Tr. 1535-36; see also G938. This fact is further confirmed by a letter from Campisi to Rogers dated May 3, 1990, in which Campisi enclosed the 20-year issue illustrations already discussed, see supra, and explained that Newport did not possess suitable issue illustrations projecting the performance of Camelot’s plan over its entire duration:

My staff is in the process of preparing a customized events oriented administration manual which will detail all of our administration procedures. In addition, we are preparing detailed financial information which will span the life of the plan.

J44, Bates 004936 (emphasis added).

Camelot attempts to justify the use of the May 1, 1990, issue illustration to analyze the economic substance of its COLI VIII plan by representing that Campisi discussed the illustrations with Rogers — presumably before Camelot purchased the plan. See Respondent’s Proposed Findings of Fact ("RPFOF”) ¶ 219; Resp. Br. at 61. This representation is flatly contradicted by the trial record, discussed above. Although the May 1, 1990, issue illustration ended up in Camelot's files at some point after May 1990, neither Rogers nor Campisi could state when. Rogers, Tr. 466-67; Campisi, Tr. 1536-37. Therefore, the preponderance of the evidence established that Camelot did not receive this illustration prior to or at the time it purchased its COLI VIII policies.

58 . Q. Now, does that refresh your recollection that possibly what happened is that Newport, in fact, was supplied with the three-scenario [May 1, 1990] issue illustration, looked at it, saw that there were large negative cash— large negative cash flows in certain years, and then decided, as you point out in your May 3rd, 1990 letter, to do something else and not to supply that particular issue illustration to Camelot?

A. That sounds reasonable.

Campisi, Tr. 1535-36.

59 . See J110, at 5 (Feb.1993 plan anniversary re-illustration) (negative after-tax cash flows in years 51 through 53 only, totaling approximately $5.5 million; pre-interest deduction cash flows are negative until year 55; total after-tax cash flows over life of policies totaled $142.9 million, significantly less than the total projected tax credits of $175.2 million, reflecting a pre-interest deduction loss of more than $32 million); Jill, at 8-11 (Feb. 1994 plan anniversary re-illustration) (revised projections eliminated negative cash flows for every plan year; interest tax deduction was projected to exceed after-tax cash flow each year up until plan year 68; the total interest deductions estimated over the life of the plan were projected to be $169.3 million, again larger than the after tax cash-flows estimated to be $127.8 over the life of the plan, reflecting a pre-interest deduction loss of more than $41 million); J125, at 11 (Feb.1995 plan anniversary re-illustration ) (no negative cash flows any year; tax deductions exceed after-tax cash flow until year 68 of plan; total tax deductions over life of the plan projected to total $169.2 million, and total after-tax cash flows estimated $127.8 million, reflecting a pre-interest deduction loss of more than $41 million); J129, at 8-11 (Feb.1996 plan anniversary re-illustration) (no negative cash flows any year; tax credits are larger than after tax cash flows for every year until the 71st plan year; total tax deductions over life of policies projected to be $163 million, compared to total after tax benefits of $121.2 million total, reflecting a pre-interest deduction loss of nearly $42 million). These ever morphing illustrations highlight the manipulability of the illustrations and further shed doubt on the correctness of the May 1, 1990 illustration.

60 . In Winn-Dixie, the Tax Court had no reason to consider employing a present value analysis since the pre-interest deduction cash flows were negative in every year of the 60-year sales illustration. See 113 T.C. at 281-84.

61 . However, nowhere in Hoag's testimony or reports did he define what constitutes a "reasonable” discount rate, nor his reasoning supporting this determination.

62 . Camelot's also submits, according to its present value calculations, that maximum positive discount rates for positive pre-tax cash flows and earnings would be 13.1479% and 13.3603%, respectively, if "grossed up.” D.I. 123, at 66. As previously stated, the Court rejects this "gross-up” analysis as improper.

63 .The Court takes judicial notice of the thirty-year maturity United States Treasury Bonds rate. See Citizens & Southern Corp. v. Comm'r, 91 T.C. 463, 503-04, 1988 WL 90987 (1988), aff'd, 919 F.2d 1492 (11th Cir.1990) (Treasury Bond rate is a fact subject to judicial notice). The annualized yield on thirty-year Treasury Bonds has never dipped below 5.58% (1998) for the period 1990 through April 2000. See FEDERAL RESERVE BULLETIN A23 (July 2000).

64 . Once Camelot began paying its premiums monthly rather than annually, its plan was administered so as to achieve zero net equity at the end of each month. Kristie Sayre, Tr. 4010-12.

65 . A small amount of unencumbered cash placed in the policies at the beginning of each policy year was fully expensed by the deduction of monthly COI charges and depleted by the end of the policy year. Sayre, Tr. 4009-10; R314/G738, at 39-41; Bossen, Tr. 2106-OS, 2110-19.

66 .Although Camelot repeatedly describes interest as a cost, because the net cost of borrowing was the same to Camelot regardless of the crediting rate/interest rate combination due to the fixed 1% spread, the choice of a higher loan interest rate did not increase Camelot's costs, but served instead to generate larger interest deductions and correspondingly greater cash flows. G1002. That a higher loan interest rate would not increase Camelot's cost of borrowing is illustrated by the fact that Camelot's cash payments to MBL to cover accrued loan interest would remain the same regardless of the selected loan interest rate. Kristie Sayre Tr., 3971-76, 4268, R314/G738 at 33-35; Hoag, Tr. 4614-17, 4619-20, R238/G740 at 12-14; G966.

67 . The COLI VIII policy was designed to preform more effectively at higher interest rates. Eisenberg, Tr. 2402; G117.

68 . Campisi explained that the COI charge "is the equivalent of a term insurance premium for the amount at risk in the policy.” Tr. 1286; see also Eisenberg, Tr. at 2189 (refer ring lo COI charges as "term insurance charges [taken] out to pay for the cost of insurance on a monthly basis, just as if you were buying term insurance”). The "amount at risk” in the policy is "the face amount less the gross cash value.” Campisi, Tr. 1286.

Additionally, Eisenberg characterized a maximally-leveraged COLI plan like Camelot’s COLI VIII as having, from a risk perspective, the characteristics of indeterminate premium annual renewable term insurance. Eisenberg, Tr. 2581-83; G527.

69 . These dividends totaled approximately $450,000. G963.

70 . The experience-rated approach used a contingency reserve that, with a one-year time lag, ensured that the total COI charges of a plan would not exceed accrued death benefits plus interest. The contingency reserve provided a mortality dividend at the end of each policy year equal to the difference between the projected death claims and actual death claims, less a “target reserve” composed of a claim lag component and a fluctuation component to cover potentially adverse deviations from the averages. In the event that a policyholder's death claims exceeded the aggregate COI charges, Hartford would recoup the COI shortfall by adding a COI surcharge up to the maximum mortality charge permitted by Form FA85. Van Etten, Tr. 2780-83, 3070-74; G597, Bates ICMG 835-837; J136; J142.

71 . Under the COLI VIII policy, the COI charges may not exceed certain maximum rates set forth in the policy. J189 at Bates 001075.

72 . In some states, the premiums subject to tax were reduced by the amount of dividends paid to the policyholder. See Bossen, Tr. 2130.

73 .Under the DAC tax, MBL/Hartford would report 7.7 percent of the policy premiums as taxable income, resulting in a tax of approximately 2.7 percent at their 35 percent federal tax rate. See Van Etten, Tr. 2725-26, 2768.

74 . The Court recognizes that the record is muddled with respect to from whose pocket the DAC tax was paid. Van Etten testified it was a component of the expense charge to Camelot. But the record also contains testimony that the DAC tax was divided equally among Camelot, MBL/Hartford and Newport. This differing testimony from Camelot witnesses need not be resolved. The DAC tax is merely an acceleration of taxable income for which MBL gets a reduction in taxable income ratably over ten years. See Van Etten, Tr. 2724-26. Moreover, to the extent MBL/Hartford determined not to pass through the DAC tax to Camelot, the economic consequence to it was de minimis.

75 . In support of its arguments, Camelot cites the declarations of two Hartford Employees, Respondent Exhibits 646 (Declaration of Mary Jane Fortin) and 647 (Declaration of David Carlson) which were offered into evidence by Camelot on May 5, 2000, prior to the closing of the evidentiary record. The Government objected to these exhibits at trial and has briefed its objections. Tr. at 4941-42; D.I. 118, at 56-57. These declarations relate to the manner in which Hartford recorded income and expenses for Camelot’s COLI VIII plan, and were to be the subject of a stipulation of the parties. While the Court is not privy to the parties discussions surrounding the stipulation, the Government maintains it objected to a stipulation proposed by Camelot based on its view that it omitted important information. Camelot apparently did not respond to an alternative proposed by the Government.

The Government objects to the Hartford declarations on the basis that they are hearsay; the declarations do not contain complete descriptions of the financial treatment of the policies and thus do not serve the purpose for which they are offered; the statements in the declarations are cumulative of the testimony of several of Respondent’s live witnesses; and accepting these declarations into evidence would undermine the Court’s March 20, 2000, Memorandum Opinion which specifically excluded the testimony of these witnesses. The Court finds the declarations to be cumulative and to add nothing beyond the testimony of Camelot’s witnesses at trial. See Bossen, Tr. 2127-28, 2130; DesRochers, Tr.1980, 1984, 2008-09; Van Etten, Tr. 2724-27, 2766-71, 2796-2802, 2803, 2934-88, 2989-3002. The Government’s objection is sustained. Because the Court does not rely on Respondents’ Exhibits 646 and 647, it need not reach the other evidentiary arguments raised by the Government.

76 . A byproduct of Ms. Sayre's analysis demonstrates that none of the cash payment of $1,207.86 could possibly have been a deductible interest expense.

77 . Stated another way by Camelot, "[t]he true economic question is what did Camelot get for borrowing and paying interest?” D.I. 123, at 60.

78 . Camelot also asserts the fact that it was required by contract and did pay off policy loans and interest out of death benefit proceeds evidences its COLI transaction had objective economic substance. The Third Circuit Court of Appeals clearly indicated that the fact that "there may be an obligation which is binding under local law is not determinative of whether” a transaction has economic substance for federal tax purposes. Wexler, 31 F.3d at 124 (quoting Weller, 270 F.2d at 298). Similarly, Camelot's contractual obligation to pay policy loans and interest is not determinative of economic substance.

79 .In its arguments related to the subjective aspects of the sham in substance analysis, Camelot also argues that the profit-potential of life insurance policies should be determined on an after tax-basis because life insurance is a tax-preferred product and/or the Court should gross-up the death benefits and cash flows in evaluating whether Camelot could reasonably expect to profit from the COLI VIII plan in the absence of the interest deductions. For the reasons stated supra, section III.A.l.a.(l)(a), the Court declines to do so.

80 . As Camelot itself points out, money is fungible. See D.I. 128 (Respondent's Reply Brief), at 21.

81 . The only document attempting such a comparison was included in a management summary prepared by the Newport group as part of a presentation package used by Cam-pisi at an April 11, 1990 Camelot meeting, which took place after Camelot had already entered into the prepayment agreement. J34, Bates S03-00241-42.

82 . This partial withdrawal was used to pay off $26 million in loan principal and $4 million in loan interest.

83 . Although this is not explicitly set forth in the Sixth Circuit Court of Appeals’ cursory opinion, the court stated it was following the reasoning of Cen-Tex and that the facts of the case were substantially identical to those of Cen-Tex and Priester. See 454 F.2d at 638 & n. 3, 639.

84 . Additionally, Camelot devotes a portion of its post-trial briefing to arguing that ACM, supra, Knetsch, supra, and Weller, supra, are distinguishable from this case. Because the Court relies on these cases only for their statements of general principles of law, and not as dispositive of this case on the facts, it need not address these arguments.

85 . To the extent the Court has addressed certain arguments in the objective economic substance section and others in the subjective business purpose section, the Court has considered all arguments with regard to both prongs of the analysis, recognizing that this is not a "rigid two-step analysis.” ACM, 157 F.3d at 248 (internal quotation omitted).

86 . Although the IRS urges the Court to follow Rose to find Camelot’s transaction to be an "abusive tax shelter,” the Court notes the Rose court in formulating its test for "generic” tax shelters explicitly chose not to use the term "abusive.” 88 T.C. at 407 n. 2.

87 . In its motion for withdrawal of reference from the bankruptcy proceeding, the Court understood the IRS to argue, as a matter of law, for purposes of determining whether a policyholder has paid the "annual premiums due” under § 264(d)(1), loading dividends and partial withdrawals must be deducted from the premium payments. See CM Holdings I, 221 B.R. at 723. However, in its post trial brief, the IRS now argues, as a matter of fact, Camelot did not pay its full annual premiums due in years four through seven because the loading dividends and partial withdrawals were not real. Either the Court misunderstood the IRS argument at the time it decided the withdrawal of reference motion or the IRS has abandoned its pure legal argument in favor of a more fact specific argument.

88 . If the interest deductions from Camelot’s COLI policy loans are "tax shelter items,” in addition to proving substantial authority for the deductions, Camelot must also prove it "reasonably believed at the time the return was filed that the tax treatment of that item was more likely than not the proper treatment.” 26 C.F.R. § 1.6662-4(g)(l)(i)(B). A "tax shelter item" is an item which "is directly or indirectly attributable to the principal purpose of a tax shelter to avoid or evade Federal income tax.” 26 C.F.R. § 1.6662-4(g)(3). Because, as discussed in the text, Camelot cannot provide substantial authority for the policy loan interest deductions, the Court need not decide whether the interest deductions were "tax shelter items” which would trigger the necessity of additional proof of reasonable belief.

89 . In addition, both the IRS and Camelot rely on opinions by various legal and tax consultants, attorneys, COLI brokers, and COLI policyholders as evidence of substantial authority or the lack thereof. However, under 26 C.F.R. § 1.6662-4(d)(3)(iii), none of this evidence is relevant in determining whether there was substantial authority for Camelot’s position.

90 . In addition, Camelot relies upon an e-mail from Teri Culbertson to Nancy Knapp, R99, a chart purporting to show the IRS's treatment of several other COLI plans offered by different carriers, R199, and the deposition lesti-mony of George Imwalle, the Issue Specialist for IRS’s COLI Program, Tr. 956-99, 3284-3324. The IRS has objected to the admission of these items of evidence as being irrelevant to the assessment of substantial authority because observations of IRS officials several years after Camelot claimed its interest deductions are not probative of whether Camelot had substantial authority at the time it filed its tax returns. The Court agrees and additionally notes that this evidence is not of the type enumerated in Treas. Reg. § 1.6662 — 4(d)(3)(iii). Accordingly, the Court finds these items of evidence not relevant to the substantial authority issue and they therefore are not considered for purposes of the Substantial Authority analysis.

91 . Camelot relies on several other aspects of its due diligence investigation which have nothing to do with assessing the tax consequences of the interest deductions. First, Camelot asserts it performed a careful study of how COLI could be used as a cost recovery vehicle for health care costs. See RPFOF ¶¶ 66-75. This includes a study of health care costs by Barbara Bald of Coopers & Lybrand. See RPFOF ¶¶ 19-26, 62, 76-78. It also includes articles by noted insurance professors and others on the use of COLI as a mechanism to fund health care liabilities. Rogers, Tr. 182-87; J19; R3. These aspects of the investigation are not relevant to Camelot’s effort to assess its risk in taking the interest deductions.

Second, Camelot relies on an opinion on the legality of COLI from its outside legal counsel, Stark & Knoll. See RPFOF ¶¶91-97. However, Stark & Knoll’s analysis was limited to whether Camelot had to notify its employees of the COLI policy, whether Camelot had an insurable interest under Ohio and Georgia state law, and the morality of insuring ones own employees without their knowledge. Knoll, Tr. 739-758. Stark & Knoll did not review the federal tax law implications of the COLI VIII policy. Knoll, Tr. 751.

In addition, Camelot relies on discussions it had with Diebold, Inc., regarding its purchase of an MBL COLI VIII policy. See RPFOF ¶¶ 107-09. However, the discussions with Diebold were centered on the issues of whether Camelot had to notify its employees of the COLI policy, whether Camelot had an insurable interest under state law, and the morality of insuring ones own employees, as well as the need, if any, to segregate funds. Rogers, Tr. 229-232 Accounting and tax questions relating to COLI were discussed but Rogers did not recall any issue which was raised. Rogers, Tr. 233.

Again, Camelot relies upon the Teri Culbertson e-mail, R99, the chart purporting to show IRS’s treatment of other COLI plans, R199, and the Imwalle testimony, Tr. 956-999, 3284-3324. The IRS again has objected to the admission of these items of evidence as being irrelevant to the assessment of reasonable cause and good faith because observations of IRS officials several years after Camelot claimed its interest deductions are not probative of whether Camelot had reasonable cause and good faith at the time it filed its tax returns. The Court again agrees this evidence is irrelevant and, as previously held, will not be considered.

92 . The Court is mindful this case arose from a bankruptcy adversary proceeding. The IRS filed a proof of claim against Camelot in the amount of $7,530,668.93 for tax years ending August 31, 1991, August 31, 1992, and August 31, 1993, and thereafter filed an amended proof of claim in the amount of $8,871,193.96 for the same tax years. J140. In addition, the IRS filed a proof of claim against CM Holdings, Inc., in the amount of $365,361.26 for the tax period ending February 28, 1994. J135. The order will be limited to the taxable years set forth in the proofs of claim, as was the Government’s proposed form of order.