Winn-Dixie Stores, Inc. and Subsidiaries v. Commissioner

113 T.C. No. 21 UNITED STATES TAX COURT WINN-DIXIE STORES, INC. AND SUBSIDIARIES, Petitioner v. COMMISSIONER OF INTERNAL REVENUE, Respondent Docket No. 5382-97. Filed October 19, 1999. P entered into a leveraged corporate-owned life insurance (COLI) program in which it purchased life insurance on approximately 36,000 of its employees and systematically borrowed against the cash value of the policies to fund the premiums. The COLI program was designed so that annual premiums, fees, and policy loan interest would exceed the projected annual death benefits and net cash value of the policies. The program was designed to generate large amounts of interest on petitioner's policy loans that petitioner intended to deduct for income tax purposes. The income tax savings from the deductions for interest and fees were projected to be substantially in excess of the projected net costs of maintaining the COLI program. In each year of operation, the COLI program projected a pretax loss and an after-tax gain. Held: P's broad-based leveraged COLI program lacked economic substance and business purpose (other than tax reduction) and is therefore a sham for tax purposes. As a result, interest on P's COLI policy - 2 - loans is not deductible interest on indebtedness within the meaning of sec. 163, I.R.C. The administrative fees associated with the COLI program are not deductible because they were incurred in furtherance of a sham. Michael J. Henke, Tegan M. Flynn, Cary D. Pugh, Thomas Crichton IV, Robert H. Cox, and Thomas P. Marinis, Jr., for petitioner. Nancy B. Herbert, Jeffrey L. Bassin, James D. Hill, and Michelle A. Missry, for respondent. RUWE, Judge: Respondent determined a deficiency of $1,599,176 in petitioner's Federal income tax for its tax year ending June 30, 1993. After concessions, the issue is whether deductions petitioner claimed for policy loan interest and administrative fees associated with certain of petitioner's corporate-owned life insurance (COLI) policies are deductible. Unless otherwise indicated, all section references are to the Internal Revenue Code in effect for the year in issue, and all Rule references are to the Tax Court Rules of Practice and Procedure. FINDINGS OF FACT Some of the facts have been stipulated and are so found. The stipulations of facts are incorporated herein by this reference. At the time the petition was filed, petitioner was a - 3 -Florida corporation with its principal office in Jacksonville, Florida. Petitioner was founded in the 1920's, and its stock is publicly traded on the New York Stock Exchange. Petitioner is a major food retailer made up of self-service food stores which sell groceries, meats, seafood, fresh produce, deli/bakery, pharmaceuticals, and general merchandise items. As of June 30, 1993, petitioner had 1,165 stores in 14 States and the Bahama Islands. Petitioner is an accrual basis taxpayer, which has adopted a 52-53 week fiscal year ending on the last Wednesday in June. Petitioner filed a consolidated corporate Federal income tax return for its fiscal year ending June 30, 1993. As of June 30, 1993, petitioner employed approximately 36,000 full-time and 69,000 part-time employees. Since 1988, all full-time employees who completed 3 months of continuous service have been eligible for a flexible benefits program called "Winn-Flex". Under Winn-Flex, employees were furnished certain benefits that they received automatically and certain optional benefits among which they could choose. Employees automatically received life insurance and accident and sickness coverage. The optional benefits included a medical plan, dental coverage, vision coverage, supplemental associate life insurance, long-term disability and two flexible spending accounts for health care and - 4 -dependent care. Petitioner self-insured the medical and life insurance benefits under Winn-Flex while the remaining benefits were insured through third parties. The life insurance coverage provided by the company under the core Winn-Flex benefit program was in effect only while a worker was a full-time employee. Petitioner provided no postretirement benefits to its employees under Winn-Flex. Early retirees covered by the Winn-Flex plan had the option of continued coverage under a separate insurance pool not paid for by petitioner. Since 1980, petitioner has also maintained a program to provide death, disability, and retirement benefits to a limited number of full-time management level employees. This program was known as the "Management Security Program" (MSP). During the fiscal year ending in 1993, 615 of petitioner's employees were covered under the MSP. In order to provide funds for specific benefits for each manager, petitioner purchased flexible premium adjustable life insurance policies on each manager (MSP policies) from American Heritage Life Insurance Co. (AHL). The MSP policies are individual policies and not group contracts. The death benefits under the individual MSP policies were tailored to cover petitioner's costs for preretirement deaths of the covered individual and to cover costs of postretirement benefits. - 5 -Petitioner's practice of purchasing MSP policies on the lives of its managers began long before 1993. During 1992 and early 1993, Wiedemann & Johnson (WJ) and The Coventry Group (Coventry) formed a joint venture (WJ/Coventry) and approached petitioner with a proposal for the purchase by petitioner of individual excess interest life insurance policies on the lives of petitioner's employees. AIG Life Insurance Company (AIG) was to be the underwriter for the proposed policies. In a letter dated January 12, 1993, Mr. Alan Buerger, chairman of Coventry, confirmed a meeting on January 14, 1993, with Mr. Richard D. McCook, petitioner's financial vice president. Included with the letter was a memorandum from Mr. Buerger and Mr. Bruce Hlavacek, chairman and chief executive officer of WJ, proposing that petitioner purchase a "broad-based COLI pool". The memorandum provided an overview section which generally described a broad-based COLI pool as consisting of a group of corporate-owned life insurance (COLI) policies covering a wide cross-section of a corporation's employees. Petitioner was the proposed beneficiary of the COLI policies to be written on the lives of petitioner's employees. WJ/Coventry's proposal focused on two issues raised by petitioner in a prior meeting. These two issues were described as "(i) achieving positive earnings in - 6 -every year; and (ii) providing an exit if the tax laws change or Winn-Dixie's appetite for interest deductions declines." The memorandum summarized the tax aspects of leveraged COLI with the following captioned diagram: Insurance Carrier IRS (1) (3) (2) Winn-Dixie - 7 - A broad-based COLI Pool applies the same principle in ways that are effective under current law. But, where each of the existing policies was designed to fund a specific executive's benefit under the MSP, the Pool that we have illustrated would cover 38,000 employees at all levels of Winn-Dixie's workforce. With respect to obtaining the employees' consent to purchase the COLI policies on their lives, the memorandum stated: We usually recommend that a company adopt or expand employee death benefits when installing a COLI Pool. This provides an immediate and meaningful benefit for employees, and it helps to provide a logic and incentive for obtaining employees' consent to being insured. The benefit may depend on the size of the pool and the amount of the insurance purchased on each employee. A death benefit in the range of $5,000 to $15,000 is typical for the Pools presented here. After an employee leaves the company, the benefit is normally reduced or discontinued. With normal rates of retirement and attrition, only a small proportion of the participants will receive a benefit. As a result, the cost of providing the benefit is insignificant. The memorandum also expressed an opinion on the tax issues raised by the proposed COLI pool, the legislative status of leveraged COLI, and exit strategies available to petitioner in the event that the tax laws change: What tax issues are raised by the COLI Pool? Deductibility of Interest. Because the COLI Pool involves systematic borrowing of increases in the policies' cash value, a deduction for interest to carry policy loans is allowed only if at least four of the first seven annual premiums are paid in cash. In addition, a deduction is allowed for interest on only the first $50,000 debt to carry policies on any one employee. The COLI Pool proposed here is designed to satisfy the 4-out-of-7 rule, and the financial illustrations take into account the $50,000 cap on loans for which interest deductions are allowed. - 8 - • * * * * * * What is the legislative status of leveraged COLI? In the past few years, Rep. Barbara Kennelly (D-Conn.) and Senator David Pryor (D-Ark.) have introduced bills that would impose new restrictions on the deductibility of interest paid on loans from COLI policies. No bill is now pending, but it is possible that one will be introduced in the future. Kennelly/Pryor, as the last such bill was generally known, was written with the participation and support of the National Association of Life Underwriters, and, if a similar bill does become law, we do not believe the financial advantages of Winn-Dixie's COLI Pool would be seriously compromised. History suggests that specific changes in the law that would address leveraged COLI would also allow grandfathering of existing policies. Past changes, for example, imposition of the $50,000 loan cap, have grandfathered existing policies, and the large number of major corporations that have created COLI pools is a significant political constituency. Of course, grandfathering cannot be assumed, and we have, therefore, kept the consequences of exit very much in mind in developing strategies for Winn-Dixie. What exit strategies are available if the tax laws or Winn-Dixie's tax position changes? A COLI Pool can become a financial burden if the tax arbitrage in the program loses its attractiveness. This can occur, for example, if Winn-Dixie's marginal tax rate on interest deductions becomes low and remains low, if Winn-Dixie becomes an alternative minimum taxpayer, or if the intended premium payment strategy becomes invalid through regulation. Likewise, Winn- Dixie's appetite for interest expense may be satisfied for reasons unrelated to deductibility. • * * * * * * If it becomes necessary or useful to terminate the COLI Pool, or to discontinue further borrowing, Winn-Dixie will be able to do so without significant adverse effect. The policies can be put on a "paid-up" basis, - 9 - either with the original carrier or with another carrier via a "1035" exchange, without incurring a tax liability, a negative effect on earnings, or a significant cash payment. The memorandum outlined two proposed financial strategies for structuring the purchase by petitioner of the pool of COLI policies. The first strategy was labeled "cash management". The second strategy was labeled the "zero-cash strategy" and was described as follows: Under Strategy 2, Winn-Dixie would maximize its tax arbitrage by borrowing the first three premiums and would minimize its cash investment by withdrawing accumulated policy values to pay the next four premiums. The policies used in this zero cash strategy are specially designed to minimize cash outflows and to maximize the rate of return on investment. Thus, loads are minimal, the interest rate is high, and the loan spread is limited to 40 basis points. Because little cash is required, a higher premium can be used. We have illustrated an average premium of $3,000 per employee. Petitioner elected strategy 2, the zero-cash strategy. On January 25 and 27, 1993, Mr. Buerger sent revised copies1 of the 1993 COLI proposal materials to Mr. McCook and Mr. Hlavacek. The revised 1993 COLI proposals outlined two scenarios for the amount of interest petitioner was to be charged for policy loans. Both interest scenarios were based upon the zero- 1 The above-quoted sections of the proposal remained substantially the same in each of the revised copies of the proposal. - 10 -cash strategy.2 The two scenarios for the zero-cash strategy were outlined as follows: Scenario 1 - Constant Loan Interest Scenario. In this scenario, it is assumed that Winn-Dixie's appetite for interest deductions remains large, and policy interest is charged at 11.06% throughout the life of the COLI Pool. This scenario results in the largest amount of earnings possible. Scenario 2 - Reducing Loan Interest Scenario. In this scenario, Winn-Dixie's appetite for interest deductions is assumed to reduce over time. To adjust to the change in circumstances, the interest rate under the COLI Pool policies is reduced to 8% after the fifteenth year. This scenario generates a somewhat smaller tax arbitrage, but the resulting earnings are nevertheless significant. Included in the revised proposal memorandum dated January 27, 1993, were projections of petitioner's profit and loss, cash-flow, and balance sheet balances under scenario 1, the constant loan interest rate scenario, for the years 1993 through 2052. The projections were based on the assumption that premiums paid by petitioner would be financed by policy loans in all years except years 4 through 7. In the years 4 through 7, inclusive, premiums were to be paid using funds from policy withdrawals. The constant loan interest rate (scenario 1) projections included in the January 27, 1993, memorandum are attached as appendix A. 2 The Jan. 27, 1993, revised proposal assumed a pool covering 38,000 employees with an average premium of $3,000 per employee. - 11 - Under the constant loan interest rate scenario, the projections assumed that the interest rate paid by petitioner on policy loans remained constant at 11.06 percent3 for the life of the pool. Amounts to be credited to petitioner on borrowed cash value were set at 40 basis points below the 11.06 percent being charged to petitioner.4 Thus, the borrowed cash value would be credited at 10.66 percent. The remaining cash value was credited at 4 percent. The January 27, 1993, memorandum explains that "A COLI Pool generally works best when the interest rate on policy loans is highest."5 The projections under the constant loan interest rate scenario were also based on the following assumptions: Corporate tax bracket 38% Population (number of insured employees) 38,000 Premium $3,000 per life Mortality assumption 100% of 1983 GAM1 Fee $8 per participant annually 1 The mortality assumption "GAM" was not defined in the proposal. Ultimately, petitioner and AIG agreed upon using the 1980 Commissioners Standard Ordinary Mortality Table B to estimate mortality. 3 The 11.06-percent rate was based on Moody's Baa rate from November 1992, which was 8.96 percent. Coventry converted it to an arrears rate and added 1 percent to reach 11.06 percent. 4 This is referred to as "loan spread". The Jan. 27, 1993, memorandum explains: "An effective COLI Pool should have a small 'spread' between the interest rate charged on policy loans and the amounts credited to borrowed cash values." 5 The rate in effect for the MSP policies was 8.41 percent. - 12 - Based upon the above assumptions, Coventry projected that the "pretax earnings effect" of the COLI plan for the first policy year (1993) would be a loss of $4,605,000,6 before adjusting for the related reduction in income taxes. This pretax loss was based on the following projected figures for the end of the first policy year: Cash surrender value (CSV) of the COLI policies of $119,586,000 increased by death benefits of $2,016,000 and reduced by annual premiums of $114 million,7 accrued loan interest of $11,902,000,8 and administration fees of $304,000.9 Similar projections for the years 1994 through 2052 produced pretax losses in each year. With respect to the estimated 1993 premiums of $114 million, the projection indicated that petitioner would, in part, satisfy the premiums by borrowing $107,684,000 against cash surrender 6 See appendix A. 7 This amount is arrived at using the assumed premium per employee of $3,000 times the assumed 38,000 employees for a total of $114 million. 8 This amount is the interest due from petitioner as calculated by Coventry on a projected first-year policy loan taken by petitioner in the amount of $107,684,000. Annual interest of 11.06 percent on $107,684,000 is actually $11,909,850. 9 Using these figures the calculation is: $119,586,000 + $2,016,000 + ($114,000,000) + ($11,902,000)+ ($304,000) = ($4,604,000). The $1,000 variance between this calculation and the pretax loss of $4,605,000 in appendix A appears to be attributable to rounding of the components of the calculation. - 13 -value. For 1994, the projection indicated that petitioner would partially satisfy its premium and interest payment10 obligation by borrowing $113,929,000 and withdrawing $1,649,000 from the net cash value of the policy. For 1995, the projection indicated that a loan of $110,318,000 and a policy withdrawal of $14,731,000 would be used to pay a large portion of the premiums of $113,852,000 and interest payment of $24,486,000. Except for the next 4 years, 1996 through 1999, the projection indicated petitioner would finance a large portion of its premium and interest payments with policy loans. For years 1996 through 1999, petitioner would generally finance its premium and interest payments through policy withdrawals. The policies' premiums were to be completely paid by 2007; therefore, no premium payments were projected for the years 2008 through 2052. However, for years 2008 through 2052, policy loans continued to be projected in amounts that were approximately between 90 percent and 95 percent of the amount of the annual loan interest payments. The projection indicated that petitioner would sustain a negative "pretax earnings effect" on the COLI plan for every year the plan remained in effect. Thus, the projection for the years 10 The projection indicated that petitioner would actually pay the estimated $11,902,000 of interest due on the 1993 loan of $107,684,000 in 1994. Interest accumulated on policy loans was payable in arrears. - 14 -1993 through 2052 indicated petitioner would incur net pretax out-of-pocket losses as follows: Pretax Pretax Pretax Loss Loss Loss Year Effect Year Effect Year Effect 1993 $4,605,000 2013 $16,390,000 2033 $15,238,000 1994 8,403,000 2014 17,839,000 2034 14,769,000 1995 11,282,000 2015 18,071,000 2035 14,320,000 1996 10,399,000 2016 18,285,000 2036 13,828,000 1997 9,997,000 2017 18,492,000 2037 13,281,000 1998 9,559,000 2018 18,546,000 2038 12,674,000 1999 9,381,000 2019 18,446,000 2039 12,013,000 2000 9,756,000 2020 18,342,000 2040 11,312,000 2001 9,959,000 2021 18,228,000 2041 10,583,000 2002 10,310,000 2022 18,103,000 2042 7,675,000 2003 10,725,000 2023 17,968,000 2043 5,867,000 2004 11,358,000 2024 17,817,000 2044 8,374,000 2005 12,215,000 2025 17,645,000 2045 7,782,000 2006 13,151,000 2026 17,446,000 2046 7,214,000 2007 14,178,000 2027 17,215,000 2047 6,650,000 2008 10,134,000 2028 16,946,000 2048 6,110,000 2009 11,269,000 2029 16,643,000 2049 5,583,000 2010 12,420,000 2030 16,193,000 2050 5,070,000 2011 13,653,000 2031 15,086,000 2051 4,586,000 2012 14,973,000 2032 16,545,000 2052 4,742,000 Total pretax loss 755,644,000 The projection of profit and loss also included an analysis of the effect of the COLI plan on petitioner's income tax liability in each of the years 1993 through 2052. Assuming a 38-percent corporate tax bracket,11 the projected $11,902,000 loan interest accrued and deducted by petitioner in the first policy 11 The corporate tax rate on all projections was an estimated combined Federal and State marginal tax rates. - 15 -year (1993) resulted in a projected tax saving of $4,524,000.12 A projected deduction for estimated administration fees of $304,000 resulted in a projected tax saving in 1993 of $116,000.13 Thus, the total projected tax benefits from deductions generated by the proposed COLI plan during the first policy year was $4,640,000.14 The projection compared the estimated tax benefit of $4,640,000 to the estimated pretax loss effect of $4,605,000 resulting in a positive "after-tax earnings effect" of $35,000.15 The after-tax earnings effect (the excess of tax savings over net pretax costs) was projected to increase substantially in subsequent years. Coventry's projected after-tax earnings effect was as follows: 12 A deduction of $11,902,000 times an assumed tax rate of 38 percent actually results in a tax benefit of $4,522,760. But see supra note 8. 13 For instance, a deduction of $304,000 times an assumed tax rate of 38 percent results in a tax benefit of $115,520. 14 Thus, $4,524,000 + $116,000 = $4,640,000. 15 For instance, a reduction in earnings due to amounts paid of $4,605,000 can be offset by a reduction in taxes of $4,640,000 for an overall after-tax earnings increase of $35,000. - 16 - After-tax After-tax After-tax Earnings Earnings Earnings Year Effect Year Effect Year Effect 1993 $35,000 2013 $60,796,000 2033 $44,780,000 1994 1,021,000 2014 59,009,000 2034 43,361,000 1995 2,770,000 2015 58,409,000 2035 41,789,000 1996 3,642,000 2016 57,797,000 2036 40,124,000 1997 4,033,000 2017 57,161,000 2037 38,369,000 1998 4,458,000 2018 56,647,000 2038 36,529,000 1999 4,624,000 2019 56,250,000 2039 34,601,000 2000 9,997,000 2020 55,819,000 2040 32,582,000 2001 16,143,000 2021 55,353,000 2041 30,479,000 2002 22,799,000 2022 54,846,000 2042 30,466,000 2003 30,108,000 2023 54,291,000 2043 29,291,000 2004 37,980,000 2024 53,683,000 2044 23,772,000 2005 46,477,000 2025 53,017,000 2045 21,361,000 2006 55,822,000 2026 52,289,000 2046 18,971,000 2007 64,479,000 2027 51,492,000 2047 16,655,000 2008 68,339,000 2028 50,620,000 2048 14,423,000 2009 66,998,000 2029 49,664,000 2049 12,313,000 2010 65,616,000 2030 48,730,000 2050 10,347,000 2011 64,127,000 2031 48,327,000 2051 8,529,000 2012 62,524,000 2032 45,233,000 2052 6,264,000 Total after-tax earnings 2,246,431,000 The projected total after-tax earnings of more than $2.2 billion were the result of total projected income tax savings of more than $3 billion less projected pretax net losses. The projected tax savings were attributable to the anticipated tax deductions for policy loan interest and fees.16 The effect on Winn-Dixie's after-tax earnings and cash-flow was projected to be positive in each year only because of tax benefits from interest and fee deductions. Absent such tax benefits, the effect on 16 Projected interest and fees over a 60-year period totaled $77,476,680,000 and $14,326,000, respectively. - 17 -Winn-Dixie's earnings and cash-flow would be negative in every year.17 In February 1993, Mr. McCook decided to have petitioner engage in the proposed broad-based COLI plan. On March 25, 1993, Mr. Hlavacek sent Mr. McCook another projection under the constant loan interest rate scenario estimating the effect of the proposed COLI plan. The projection estimated, among other things, the effect over 60 years beginning in 1993 of the proposed COLI purchase on petitioner's effective tax rate. The projection assumed an effective tax rate of 40 percent and predicted that the proposed COLI purchase would reduce petitioner's effective tax rate each year, reaching its lowest point of 26.54 percent in 2007. AIG participated in the development of information for projections regarding petitioner's proposed COLI plan. A preliminary census reflecting the ages of the approximately 36,000 employees to be insured was prepared on May 28, 1993.18 Mr. Buerger sent two more sets of revised projections to Mr. McCook. The first set of revised projections was sent on May 28, 17 Similarly, the proposal memorandum projected the effect of the COLI purchase on petitioner's after-tax retained earnings balance on its balance sheet over 60 years. The proposal predicted that petitioner's retained earnings balance would increase by $2,241,491,000 over the 60 years. 18 Before the preliminary census, the projections were based on estimates of the number of employees and their ages. - 18 -1993, and the second was sent on June 4, 1993. Both projections were based on issuance of the proposed policies effective as of March 1993. The first set of revised projections assumed that petitioner's taxes were paid at a rate of 40 percent19 for purposes of predicting the tax effect of the COLI purchase on petitioner's financial ratios. The second set of revised projections used a tax rate of 39 percent20 and assumed just over 36,000 of petitioner's employees would be insured. This projection also assumed that maximum loans would be made in most years but that cash withdrawals would be made in policy years 4 through 7 and policy years 16 through 21. The June 4, 1993, projections are attached as appendix B. Using the same basic analysis that had been used in the January projections, the June 4, 1993, projections again indicated that petitioner would sustain a pretax loss and a posttax profit on the COLI policies for every year the plan remained in effect. See appendix B. Thus, for the years 1993 through 2052, the June 4, 1993, projections indicated the broad-based COLI plan would affect petitioner's profit and loss as follows: 19 See supra note 11. 20 See supra note 11. - 19 - Pretax Tax Benefit Tax Benefit After-tax Policy Earnings From Interest From Admin Fee Earnings Year1 Effect Deduction Deduction Effect3 2 1993 ($4,188,000) $4,368,000 $113,000 = $292,000 1994 (7,885,000) 8,988,000 113,000 = 1,217,000 1995 (10,869,000) 13,887,000 113,000 = 3,130,000 1996 (9,972,000) 13,856,000 112,000 = 3,996,000 1997 (9,669,000) 13,823,000 112,000 = 4,266,000 1998 (9,366,000) 13,788,000 112,000 = 4,533,000 1999 (9,063,000) 13,750,000 111,000 = 4,799,000 2000 (9,243,000) 19,159,000 111,000 = 10,027,000 2001 (9,736,000) 25,088,000 111,000 = 15,462,000 2002 (10,013,000) 31,595,000 110,000 = 21,692,000 2003 (10,345,000) 38,733,000 110,000 = 28,498,000 2004 (10,879,000) 46,551,000 110,000 = 35,782,000 2005 (11,619,000) 55,104,000 109,000 = 43,595,000 2006 (12,428,000) 64,456,000 109,000 = 52,136,000 2007 (13,208,000) 73,436,000 108,000 = 60,336,000 2008 (9,182,000) 73,039,000 107,000 = 63,965,000 2009 (9,409,000) 72,612,000 107,000 = 63,310,000 2010 (9,651,000) 72,153,000 106,000 = 62,608,000 2011 (9,888,000) 71,661,000 105,000 = 61,878,000 2012 (10,104,000) 71,134,000 105,000 = 61,134,000 2013 (10,056,000) 70,569,000 104,000 = 60,616,000 2014 (11,085,000) 70,497,000 103,000 = 59,515,000 2015 (11,875,000) 69,905,000 102,000 = 58,132,000 2016 (12,704,000) 69,181,000 101,000 = 56,579,000 2017 (13,584,000) 68,405,000 100,000 = 54,921,000 2018 (14,384,000) 67,581,000 99,000 = 53,296,000 2019 (15,112,000) 66,710,000 98,000 = 51,696,000 2020 (15,905,000) 65,789,000 96,000 = 49,980,000 2021 (16,447,000) 64,818,000 95,000 = 48,467,000 2022 (16,336,000) 63,797,000 94,000 = 47,554,000 2023 (16,124,000) 62,724,000 92,000 = 46,692,000 2024 (15,898,000) 61,599,000 91,000 = 45,791,000 2025 (15,662,000) 60,421,000 89,000 = 44,848,000 2026 (15,417,000) 59,191,000 87,000 = 43,862,000 2027 (15,163,000) 57,908,000 85,000 = 42,831,000 2028 (14,901,000) 56,573,000 84,000 = 41,756,000 2029 (14,632,000) 55,187,000 82,000 = 40,637,000 2030 (14,352,000) 53,750,000 80,000 = 39,478,000 2031 (14,063,000) 52,264,000 78,000 = 38,279,000 2032 (13,765,000) 50,732,000 75,000 = 37,043,000 2033 (13,457,000) 49,155,000 73,000 = 35,771,000 2034 (13,141,000) 47,535,000 71,000 = 34,465,000 2035 (12,816,000) 45,873,000 69,000 = 33,126,000 2036 (12,483,000) 44,172,000 66,000 = 31,755,000 2037 (12,139,000) 42,435,000 64,000 = 30,360,000 2038 (11,784,000) 40,665,000 61,000 = 28,942,000 2039 (11,417,000) 38,865,000 59,000 = 27,507,000 2040 (11,035,000) 37,040,000 56,000 = 26,060,000 2041 (10,638,000) 35,194,000 53,000 = 24,609,000 2042 (10,223,000) 33,332,000 51,000 = 23,160,000 2043 (9,794,000) 31,460,000 48,000 = 21,714,000 2044 (9,344,000) 29,583,000 45,000 = 20,284,000 2045 (8,902,000) 27,707,000 43,000 = 18,847,000 2046 (8,455,000) 25,840,000 40,000 = 17,425,000 2047 (8,006,000) 23,990,000 37,000 = 16,022,000 2048 (7,563,000) 22,167,000 34,000 = 14,638,000 2049 (7,153,000) 20,378,000 32,000 = 13,256,000 2050 (6,739,000) 18,630,000 29,000 = 11,921,000 2051 (6,407,000) 16,932,000 27,000 = 10,552,000 2052 (6,244,000) 15,292,000 24,000 = 9,072,000 (681,922,000) ?? ?? = ?? 1 Policy years are from Mar. 1 to the end of February. 2 Based on the underlying figures from which this amount was computed, the figure should be $4,189,000. 3 In some instances the figures in this column vary from the sum of their components by $1,000. This appears to be due to rounding off, and any variances appear to be insignificant. - 20 - The June 4 projections indicate that without tax savings from policy loan interest and administration fee deductions, the earnings effect over 60 years would have been a negative $681,922,000. The tax savings over 60 years for interest and fee deductions were projected to be $2,696,038,000. After these tax benefits are taken into consideration, the projection indicated that petitioner would realize an after-tax profit of $2,014,115,000. These after-tax financial benefits were dependent on the tax savings flowing from the interest and fee deductions being generated by the broad-based COLI plan. Petitioner prepared a list of "Company Expense Reduction Opportunities" for fiscal year 1993 that listed 23 items of savings that totaled $329,093,000. The largest item of expense reduction on the list is "Proposed Corporate Life Insurance (COLI)". The list shows that this item (COLI) was estimated to be implemented on June 30, 1993, and that petitioner's estimated savings from COLI was $300 million. Mr. McCook testified that he thought this figure was derived from the "After-Tax Earnings Effect" shown in column J of appendix A.21 AIG is a major underwriter of COLI policies. AIG had been working with WJ/Coventry in preparing the COLI plan. Mr. Qureshi 21 Three hundred million dollars is the approximate total of the annual "After-Tax Earnings" projected for the policy years 1993 through 2007. The total projected After-Tax Earnings for the years 1993 through 2052 is $2,246,431,000. See appendix A. - 21 -of AIG was the actuary who designed and priced the COLI policies purchased by petitioner. On June 4, 1993, Mr. Larry Walters of AIG, sent a Letter of Understanding regarding the COLI policies to Mr. McCook. The Letter of Understanding provided that petitioner would remit, as consideration for the policies, a net payment of $7,245,000, the difference between a total premium of $108,573,000 and an estimated first year policy loan of $101,328,000. In the Letter of Understanding, AIG agreed to provide life insurance coverage in accordance with the Policy Terms Overview. The Letter of Understanding required that the following conditions be met: 1. * * *[Petitioner] reviews and verifies the accuracy of the information contained on the Client Master Information Form, attached as Exhibit A.[22] * * * 2. * * *[Petitioner] completes and certifies the information contained on the Certification of Employee Census form, attached as Exhibit B,[23] 22 Exhibit A, Client Master Information Form, listed petitioner's name as the name to appear on the policy and as the owner of the policy. Petitioner's main address was listed as the billing address, and Mr. McCook was named as the contact. The policy name was listed as "Excess Interest Life Ins.", and the effective date was listed as Mar. 1, 1993. 23 Exhibit B, Certification of Employee Census, generally required that petitioner, as part of the application for coverage under the policies, certify that employee information on the census was correct. Among other things, petitioner certified that each individual on the census was a full-time (minimum 30 hours per week) employee, at least 18 years of age, no older than age 75, not absent from work for more than 10 consecutive business days within the 90 days preceding the date of - 22 - • * * 3. * * *[Petitioner] completes at least one Individual Life Insurance Application for each defined class of covered employees. 4. The aggregate number of employees within each defined class of covered employees, on the effective date of coverage, totals at least 2,000 lives. 5. The minimum annual premium for each covered employee is $3,000 and the maximum annual premium for each covered employee is $16,667. 6. AIG Life determines that each defined class of covered employees identified on the attached Client Master Information Form and each employee identified on the Certification of Employee Census are acceptable for underwriting purposes and that the amounts of coverage applied for are acceptable. In addition, the Letter of Understanding required that petitioner acknowledge and agree with the following: 1. This Letter of Understanding and attached Exhibits A and B contain the entire agreement between the parties and supersedes all previous agreements entered into between Client and AIG Life, or promises made with regard to the subject matter of this letter. 2. * * * [Petitioner] has reviewed with its own legal and tax advisors all present and future implications of its ownership of the Policies, including, but not limited to, the tax consequences of loans and/or withdrawals from the Policies and the deductibility thereof, and that it has not relied upon any representations of AIG certification, and that petitioner notify and obtain consent from each employee on the census that insurance is to be issued on his or her life. - 23 - Life or any employee, broker, or agent of AIG Life in that regard. 3. The premiums specified in the Policies are intended to meet the requirements of Section 7702 and Section 7702A of the Internal Revenue Code as in effect on the date of this letter so that the Policies will qualify as life insurance and will not be treated as modified endowment contracts. AIG Life does not warrant or represent that the Policies will not be treated as modified endowment contracts. 4. The statements contained in the attached Exhibits A and B shall be considered as binding representations by the * * * [petitioner] and that such Exhibits shall be deemed attached to and made a part of the Policies. The Letter of Understanding provided that if, within 90 days of June 4, 1993, all the above conditions had been met and AIG received the total first-year premium, then AIG would agree to issue to petitioner, in the State of Florida, the life insurance policies effective as of March 1, 1993. By invoice dated June 9, 1993, Coventry requested payment of a balance due on the 1993 COLI policies of $7,245,000 from petitioner. The invoice reflected a total premium of $108,573,000, covering 36,191 total lives at $3,000 of premium each, less a policy loan of $101,328,000 to arrive at the net amount due. On June 15, 1993, in accordance with the Letter of Understanding, AIG sent the "Policy Terms Overview" (PTO) to petitioner. The PTO provided for an effective date of March 1, - 24 -1993, and required that AIG and petitioner agree to several provisions under the insurance policies relating to the following: Claim Stabilization Reserve, Cost of Insurance Rates, Expense Caps, Surrender Fee, Interest Rate on Unborrowed Funds, and the Loan Interest Spread. The PTO generally provided that AIG would establish on behalf of petitioner a claims stabilization reserve (CSR) for the policies. Petitioner could not withdraw or borrow against the amounts credited to the CSR. The maximum level of the CSR at the end of each year was generally determined to be the higher of the annualized "cost of insurance"24 (COI) charges actually collected in any one of 3 preceding policy years or the highest amount of death benefits actually incurred in any one of 3 policy years. COI charges were deducted from the premium that was paid and used to fund the CSR. The CSR was generally held available by AIG to pay death claims under the COLI policies. 24 "Cost of Insurance" was defined in the policy document. The COI was calculated on each monthly processing date. Generally, the COI was calculated under the following formula: COI = (Proceeds - Account Value) x (Value from Table of 1,000 Maximum Insurance Rates) Proceeds were defined as the benefits due to petitioner as the beneficiary. Account Value on the policy date was defined as the initial net premium less an annual expense charge. - 25 - On each policy anniversary, the PTO required AIG to compute an "Experience Cost", defined as COI charges collected from petitioner during the preceding year, less a 2-percent AIG retention fee, less the net amount for which AIG was at risk for claims during the preceding policy year. The negotiated 2-percent retention fee limited AIG's mortality related profit. If the experience cost as of a policy anniversary was positive, it was added to the CSR's value as of the policy anniversary. If the experience cost was negative, it was subtracted from the CSR's value. AIG credited interest to the CSR at an annual rate of 4 percent. If, on a policy anniversary, the CSR balance exceeded its maximum permissible level, the excess was credited to the unrestricted policy account value. The PTO provided that the CSR would be held by AIG on behalf of petitioner for as long as the policies remained in force plus 1 year. At the end of the 1 year, a final accounting would be made by AIG and the balance of any remaining reserve would be refunded to petitioner. Petitioner was to be charged 11.06 percent interest on amounts that it borrowed against the cash value of the policies. Pursuant to the PTO, the portion of the policy account value that was borrowed would earn interest at a rate not to exceed 40 basis points25 or four-tenths of 1 percent below the amount charged on 25 Each basis point equals one one-hundredth of a percent (0.01 percent). Thus, 40 basis points is equivalent to 0.40 - 26 -policy loans. Thus, the interest rate credited on the portion of the account value that had been borrowed was 10.66 percent.26 The balance of the policy account values would earn interest at a rate guaranteed to be no less than 4 percent. The policies provided for expense charges which would not exceed 23 percent of the premiums paid. The expense charges were comprised of premium expense charges of 17.8 percent and annual expense charges of 5.2 percent. The negotiated PTO reduced maximum expense charges from 23 percent to 8.934 percent of premiums paid. Mr. McCook approved the purchase of the 1993 COLI policies, and on June 17, 1993, Mr. McCook, on behalf of petitioner, executed the June 4, 1993, Letter of Understanding. On the same day, petitioner remitted the net amount of $7,245,000 to AIG as payment for the policies. The 1993 COLI policy forms were registered and approved in form and for sale by the insurance commissioner of the State of Florida. Initially, the total number of lives included, subject to eligibility, under the 1993 COLI policies was 36,191 as of June 17, 1993. percent. 26 The 10.66 percent figure is arrived at by reducing the loan interest rate of 11.06 percent by 0.40 percent. - 27 - On June 17, 1993, petitioner paid WJ/Coventry $300,000 pursuant to an invoice dated June 11, 1993. The payment was for services to be performed in year 1 of an administration agreement for the 1993 COLI policies. Also on June 17, 1993, Mr. McCook signed a Notification Certificate as a condition precedent to the sale of the COLI policies in which petitioner certified that it would notify the employees of the purchase of the COLI policies and give them an opportunity to refuse the coverage. On July 19, 1993, Mr. McCook executed the PTO.27 Also on July 19, 1993, Mr. McCook accepted receipt of the COLI policy contract documents from AIG and authorized Coventry to retain possession of the policies on petitioner's behalf.28 The type of coverage provided was listed as "excess interest life". The person whose life was insured under each policy was one of petitioner's employees. Petitioner was listed as the owner and beneficiary of each policy. The rights and liabilities of AIG and petitioner were governed by insurance policy forms (Policy Form), riders to the policies, the Letter of Understanding, and the PTO. The Letter of Understanding and the PTO set forth essential elements of the 27 Mr. Walters of AIG signed the PTO on July 15, 1993. 28 The contracts were listed on the delivery receipt as being policy Nos. 5003000001 through 5003035983. - 28 -agreement between petitioner and AIG and amended and tailored the COLI policies to petitioner. Policy face amounts varied with the age of each insured. Generally, petitioner's death benefits were governed by policy Option A, which provided for benefits based on the larger of the face amount plus the account value on the date of death, or the account value on the date of death multiplied by a specified percentage based on the age of the insured.29 The applicable mortality table was the 1980 Commissioners Standard Ordinary Mortality Table B, Age Last Birthday, referred to as the CSO-B table. Under the terms of the policies, death benefits from a policy would first be used to reduce any outstanding loan. Under the terms of the policies, petitioner could withdraw part of the cash value of each COLI policy. However, a withdrawal could not exceed the "net cash value"30 of the policy. The policies also permitted petitioner to borrow an amount that, with interest to the next policy anniversary, would not 29 Some of the death benefits were paid under Option B, which was calculated to include the account value in the face amount, and the insurance proceeds were determined to be the larger of the face amount on the date of death, or the account value on the date of death multiplied by a specified percentage. 30 The net cash value of the policy was defined as the cash value less any prior withdrawals and any policy debt. The cash value was defined as the greater of the Guaranteed Cash Value, or the Account Value less the surrender charge that applies. The Guaranteed Cash Value was determined by a Table of Guaranteed Values provided with the policy. - 29 -exceed the net cash value of the policy. The policies provided that at the insured's death, any policy debt will be deducted from the proceeds of that policy. The policies were the sole security for any loans. Interest on petitioner's loans was due on each policy anniversary date. The policies provided for both a fixed and variable loan interest rate. The policies were modified by a Renewable Level Term Insurance Rider. The rider provided death benefits equal to the amount of death benefits lost as a result of a withdrawal from the account value. Under the provisions of the rider, petitioner had the option on any policy anniversary date to elect to change from the 11.06-percent policy loan interest rate to a fixed rate of 10 percent in arrears or 9.1 percent in advance which would apply to both old and new policy loans. For the 1993 COLI policy year March 1, 1993 to 1994, petitioner elected the variable loan interest rate of 11.06 percent. The final provisions governing the COLI policies purchased by petitioner in June 1993 were devised to produce results in accord with those set forth in the June 4, 1993, projections contained in appendix B. On July 19, 1993, petitioner entered into an Administrative Services Agreement with Coventry in which petitioner appointed Coventry as the administrator of the COLI pool. Under the - 30 -agreement, petitioner was to pay $300,00031 in each of the first 2 years and $200,000 annually in all subsequent policy years, and Coventry was to perform the following services in connection with the COLI policies: (a) On the basis of census information supplied by petitioner (1) Identify which of the petitioner's employees may become insured; (2) Calculate the face amounts of the policies and the first year premiums for the covered employees; (3) Determine and process new insureds and process any change in the status of any insured; (4) Make the necessary calculations with respect to premiums, loans, withdrawals, loan interest, death claims and/or any other periodic payments; (b) Provide consolidated invoice and itemization to petitioner and AIG; (c) Receive and inspect the insurance policies from AIG and forward them to petitioner; (d) Search government databases and other sources for covered deceased employees and obtain death certificates for deceased insureds; (e) Provide petitioner with ongoing advice with respect to financial options and strategies related to 1993 COLI polices; and (f) Provide petitioner with various reports including insurance value reports, year-end summaries, accounting reports and custom-designed decision- support reports. 31 Petitioner deducted $100,000 of the $300,000 on its income tax return for the fiscal period ending June 30, 1993. - 31 - For employees who died while in petitioner's employ, petitioner filed an Employer's Statement through its plan administrator, Coventry, who would then present the claim to AIG. Coventry, as plan administrator, was responsible for ascertaining employee deaths for those employees that died while no longer in petitioner's employ. Westport Management, an organization that performs administrative services for life insurance companies, was engaged by AIG and WJ/Coventry to administer petitioner's COLI policies. In order to ascertain deaths of former employees, Coventry would ask Westport to perform "Social Security sweeps", by checking data base files to determine whether any covered former employee had died. After the purchase of the policies and input of the COLI policy data on its computer system, Westport began its administrative duties, regularly preparing performance and accounting reports, which were provided to Coventry and AIG. On every policy anniversary, Westport calculated all values on a monthly basis for all policies for the coming year. On petitioner's behalf, Coventry administered the COLI plan, acted as an intermediary with AIG, and checked reports and other information provided by AIG and Westport to ensure correctness. Every month and on request, Coventry received reports from Westport on past and expected performance of the policies and policy loans. From these reports, Coventry generated annual and - 32 -periodic policy value and other reports and journal entries showing aggregate policy activity. The Coventry reports included information about the CSR, experience rating, cash value calculations, claims, refunds and recisions. The Coventry reports also included information about the amount of tax savings the COLI program was generating for petitioner. From 1993 through 1996, petitioner kept the employee COLI policies in force. AIG billed petitioner for premiums and interest annually on a net basis, as set forth below: June 1993, Policy year beginning March 1, 1993: Premium $108,573,000 Loan (101,328,000) 1 Net premium 7,245,000 1 A revised invoice was sent on Sept. 21, 1993, which reflected a reduction in insureds from 36,191 to 35,983 due to a recision of 208 policies. Thus, the calculation was as follows: Premium $107,949,000.00 Loan (100,770,140.06) Net premium 7,178,859.94 Less amount paid (7,245,000.00) Balance owed petitioner 66,140.06 Policy year beginning March 1, 1994: Premium $107,862,000.00 Loan (108,877,159.95) Interest 11,136,375.63 Balance due 10,121,215.68 - 33 - Policy year beginning March 1, 1995: Premium $107,685,000.00 Loan (112,165,202.89) Withdrawal (4,080,660.74) Net premium due (8,560,863.63) Interest 23,140,858.57 1 Balance due 14,579,994.94 1 Policy year beginning Mar. 1, 1995, was revised at least twice. The final revision resulted in the following: Premium $107,685,000.00 Loan (112,112,913.04) Withdrawal (4,134,020.12) Net premium due (8,561,933.16) Interest 23,140,858.57 Balance 14,578,925.41 Amount paid 14,579,994.94 Net refund due 1,069.53 Policy year beginning March 1, 1996: Premium $107,553,000.00 Withdrawal (129,934,414.41) Net premium due (22,381,414.41) Interest 35,497,690.97 Balance due 13,116,276.56 COI and policy expense charges (DAC tax, State premium tax, commission and loading charges) under petitioner's COLI policies were as follows: Policy Cost of Expense Year Insurance Charges Total 1993 $3,354,561 $3,412,447 $6,767,008 1994 4,641,249 4,721,130 9,362,379 1995 4,890,649 8,516,817 13,407,466 1996 5,173,414 8,990,642 14,164,056 The annual amounts of cash paid by petitioner to AIG for the COLI policies, compared with the total of COI and policy expense charges for corresponding years, were as follows: - 34 - COI Plus Policy Cash Paid By Expense Year Petitioner Charges 1993 $7,178,860 $6,767,008 1994 10,121,216 9,362,379 1995 14,578,925 13,407,466 1996 13,116,277 14,164,056 Total 44,995,278 43,700,909 For the first 4 COLI policy years, beginning March 1, 1993, AIG billed petitioner $431,049,000 in gross premiums and $69,774,925 in interest charges. Gross premiums and interest charges totaled $500,823,925. Of this total amount, petitioner remitted the above-calculated $44,995,27832 in cash. Following the enactment of tax law changes in 1996, petitioner commenced discussions with AIG, Coventry, and WJ concerning the phaseout or discontinuance (unwind) of the COLI policies. These discussions concerned the approximately 36,000 policies purchased by petitioner in 1993 plus approximately 11,000 and 9,000 COLI policies purchased in 1994 and 1995, respectively. On October 22, 1996, Mr. McCook sent a letter to Mr. Qureshi, vice president of AIG, which stated: Winn-Dixie has used AIG policies on three corporate owned life insurance (COLI) programs over the last several years. Because of the recent tax law changes, we have been working with Alan Buerger of the Coventry Group and Bruce Hlavacek of Wiedemann and Johnson, to try to minimize the financial impact on Winn-Dixie for the phase out of COLI. Mr. Qureshi responded to Mr. McCook's October 22, 1996, letter 32 See preceding table for cash total. - 35 -and acknowledged the recent change in the law with respect to the COLI policies. Mr. Qureshi also indicated that AIG would be pleased to discuss various options available to petitioner. Coventry prepared a draft booklet dated October 30, 1996, which contained, among other things, an overview of the current status of petitioner's COLI pool, an opinion of the financial effect of the 1996 tax law change, and explanations of several exit and unwind strategies. The draft booklet indicated that petitioner had three separate enrollments covering approximately 55,740 lives. The first enrollment "WD1" was in relation to the policies written in 1993 covering 35,810 employees. The second enrollment "WD2" was in relation to the policies written on November 30, 1994, covering 10,704 employees. The third enrollment "WD3" was written on June 30, 1995, and covered 9,226 employees. With respect to the effect of the 1996 tax law changes on petitioner's COLI policies, the booklet stated in pertinent part: In August of 1996, Congress amended the Internal Revenue Code was [sic] to deny deductions for any interest on policy loans on the lives of employees, officers, and persons financially interested in a trade or business maintained by the taxpayer. The disallowance was retroactive to January 1, 1996, except that deductions may be continued through 1998 on up to 20,000 policies. The deduction on those policies, however, must be based on an interest rate no higher than Moody's average corporate bond rate, and only 90% - 36 - of such interest is deductible in 1997 and 80% thereof is deductible in 1998. • * * * * * * In the aggregate, the three enrollments cover 55,740 lives with aggregate outstanding loans of about $500 million at interest rates averaging 11%. At a 39% tax bracket, these policies would produce tax deductions worth $21,450,000 per year. Under the amended law, assuming aggregate indebtedness of $195 million on the "best" 20,000 policies and a Moody's rate of 8% per annum, the following savings will be available: Calendar 1996 $6,084,000 Calendar 1997 5,475,600 Calendar 1998 4,867,200 The booklet next identified three basic exit strategies for petitioner. The three strategies were listed as the policy surrender, policy unwind, and aggressive tax strategy. The policy surrender strategy generally entailed the cancellation or surrender of the policy and the receipt by petitioner of the net cash value of the policy. The booklet recommended under this strategy that petitioner maintain the policies on 20,000 lives in fiscal years 1996 and 1997. With respect to the policy unwind strategy, in lieu of surrendering the policies, petitioner was informed that it could keep the policies in force and allow the unrealized gains related to the policies to be paid out eventually as tax-free death benefits. The booklet further stated that in order to unwind a policy, petitioner "would withdraw a portion of the cash value - 37 -equal to premiums paid (i.e., Winn-Dixie's tax basis) and apply the withdrawal to repay an equal amount of loan." The booklet indicated that the result of such a withdrawal and repayment is a policy with a greatly reduced cash value, substantially all of it borrowed. The third strategy, the aggressive tax strategy, suggested that under the revised statute, deductions were disallowed only with respect to policies on the life of an individual who was an officer or employee or was financially interested in petitioner's trade or business. The booklet further indicated that counsel for Coventry believed that a strong argument could be made that the disallowance described by the statute did not apply where the insured was a former officer or employee or was not financially interested. Based on this argument, the booklet gave an example which assumed an additional $200 million of aggregate indebtedness could be attributed to petitioner's former employees upon whom policies were still maintained. As a result of the additional $200 million of aggregate indebtedness, the booklet concluded that the tax savings in each year would be equal to 39 percent of 11 percent of $200 million or $8,580,000, for as long as the loans remained in force. In a letter to Mr. Qureshi dated September 8, 1997, Mr. McCook indicated that in light of the passage of the legislation pertaining to leveraged COLI, petitioner was working toward a - 38 -more complete understanding of the COLI policies it purchased from AIG. Finally, in letters dated December 4, 1997, Mr. McCook notified Mr. Qureshi and Mr. Buerger of petitioner's intent to cancel all three blocks of leveraged COLI policies. Mr. McCook indicated in his notice to Mr. Qureshi that petitioner wished to surrender COLI blocks I, II, and III as of November 1, October 30, and June 30, 1997, respectively. OPINION On its return for the fiscal year ending June 30, 1993, petitioner claimed a deduction of $3,735,544 for accrued interest on loans from COLI policies that petitioner purchased in 1993.33 Petitioner also claimed a $100,000 deduction for administrative fees related to these COLI policies.34 Respondent disallowed the deductions after determining that the 1993 COLI Plan was tax motivated, unsupported by any independent business purpose, and 33 This was approximately one-third of the total policy loan interest that would accrue during the first policy year that began Mar. 1, 1993, and ended Feb. 28, 1994. The $3,735,544 was interest attributable to the period Mar. 1 through June 30, 1993. We note that petitioner deducted interest on these "loans" for the period Mar. 1 through June 30, 1993, even though the COLI policies and policy loans were not finalized until mid-June 1993. Respondent argues that interest cannot accrue for a period prior to the time the loan was actually made. Because of our disposition, we need not address this issue. 34 This was one-third of the $300,000 administrative fee for the first policy year that began Mar. 1, 1993, and ended Feb. 28, 1994. - 39 -lacked economic substance. Respondent argues that the arrangement was a sham. The starting point for determining whether the form of a particular transaction will be recognized for tax purposes is the Supreme Court's decision in Gregory v. Helvering, 293 U.S. 465, 469 (1935), wherein the Court stated: The legal right of a taxpayer to decrease the amount of what otherwise would be his taxes, or altogether avoid them, by means which the law permits, cannot be doubted. * * * But the question for determination is whether what was done, apart from the tax motive, was the thing which the statute intended. In Gregory, the Court denied reorganization treatment with respect to a stock distribution even though the taxpayers had followed each step required by the Code for a reorganization. In deciding that the distribution was taxable as a dividend, the Court held that the structure of 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. Because the transaction lacked economic substance, as opposed to formal reality, it was not "the thing which the statute intended." Id.; see Kirchman v. Commissioner, 862 F.2d 1486, 1490-1491 (11th Cir. 1989), affg. Glass v. Commissioner, 87 T.C. 1087 (1986). A transaction that lacks substance is not recognized for Federal tax purposes. See ACM Partnership v. Commissioner, 157 - 40 -F.3d 231, 247 (3d Cir. 1998), affg. in part and revg. in part T.C. Memo. 1997-115; United States v. Wexler, 31 F.3d 117, 122 (3d Cir. 1994). Denial of recognition means that such a transaction cannot be the basis for a deductible expense. See United States v. Wexler, supra at 122. Citing the Supreme Court's decision in Gregory, the Court of Appeals for the Eleventh Circuit in Kirchman v. Commissioner, supra, stated the doctrine as follows: The sham transaction doctrine requires courts and the Commissioner to look beyond the form of a transaction and to determine whether its substance is of such a nature that expenses or losses incurred in connection with it are deductible under an applicable section of the Internal Revenue Code. If a transaction's form complies with the Code's requirements for deductibility, but the transaction lacks the factual or economic substance that form represents, then expenses or losses incurred in connection with the transaction are not deductible. [Id. at 1490.] Because the transactional events at issue in this case actually occurred, we limit our inquiry to the question of whether the substance of the COLI transaction corresponds with its form.35 35 In Kirchman v. Commissioner, 862 F.2d 1486, 1492 (11th Cir. 1989), affg. Glass v. Commissioner, 87 T.C. 1087 (1986), the court observed: Courts have recognized two basic types of sham transactions. Shams in fact are transactions that never occur. In such shams, taxpayers claim deductions for transactions that have been created on paper but which never took place. Shams in substance are transactions that actually occurred but which lack the - 41 - Section 163(a) provides that "There shall be allowed as a deduction all interest paid or accrued within the taxable year on indebtedness." Court opinions have clearly established that a lack of economic substance may operate to bar interest deductions arising under section 163. See Knetsch v. United States, 364 U.S. 361 (1960);36 United States v. Wexler, supra; Goldstein v. Commissioner, 364 F.2d 734 (2d Cir. 1966), affg. 44 T.C. 284 (1965). Interest payments are not deductible if they arise from transactions "that can not with reason be said to have purpose, substance, or utility apart from their anticipated tax consequences." Goldstein v. Commissioner, supra at 740; see also Sheldon v. Commissioner, 94 T.C. 738 (1990). "Such transactions are said to lack 'economic substance.'" Lee v. Commissioner, 155 F.3d 584, 586 (2d Cir. 1998) (quoting Jacobson v. Commissioner, 915 F.2d 832, 837 (2d Cir. 1990), affg. in part and revg. in part T.C. Memo. 1988-341), affg. in part and remanding in part on another ground T.C. Memo. 1997-172. The fact that an enforceable debt exists between the substance their form represents. * * * 36 In Knetsch v. United States, 364 U.S. 361 (1960), the Court applied sec. 163(a) of the 1954 Code. The language of sec. 163(a) of the 1954 Code remained unchanged in the 1986 Code. See sec. 163(a); see also United States v. Wexler, 31 F.3d 117, 123 (3d Cir. 1994). - 42 -borrower and lender is not dispositive of whether interest arising from that debt is deductible under section 163. Rather, the overall transaction, of which the debt is a part, must have economic substance before interest can be deducted. See Lee v. Commissioner, supra at 587; United States v. Wexler, supra at 125. If this were not the rule, every tax shelter, no matter how transparently sham, could qualify for an interest expense deduction as long as there was a real creditor in the transaction that demanded repayment. Such a result would be "contrary to the longstanding jurisprudence of sham shelters from Knetsch on down." Lee v. Commissioner, supra at 587. In determining whether a transaction or series of related transactions constitute a substantive sham, both this Court and a majority of the Courts of Appeals have utilized a flexible analysis that focuses on two related factors, economic substance apart from tax consequences, and business purpose. See ACM Partnership v. Commissioner, supra; Karr v. Commissioner, 924 F.2d 1018, 1023 (11th Cir. 1991); accord Casebeer v. Commissioner, 909 F.2d 1360 (9th Cir. 1990), affg. in part and revg. in part on another ground Larsen v. Commissioner, 89 T.C. 1229 (1987); James v. Commissioner, 899 F.2d 905, 908-909 (10th Cir. 1990), affg. 87 T.C. 905 (1986); Shriver v. Commissioner, 899 F.2d 724, 727 (8th Cir. 1990), affg. T.C. Memo. 1987-627; - 43 -Rose v. Commissioner, 868 F.2d 851, 854 (6th Cir. 1989), affg. 88 T.C. 386 (1987); Kirchman v. Commissioner, supra; United Parcel Serv. of Am., Inc. v. Commissioner, T.C. Memo. 1999-268.37 Economic substance, in this context, is determined by objective evaluation of changes in economic position of the taxpayer (economic effects) aside from tax benefits. See Kirchman v. Commissioner, supra at 1492; accord Knetsch v. United States, supra at 366 ("nothing of substance to be realized * * * from this transaction beyond a tax deduction"); ACM Partnership v. Commissioner, supra at 248; Sheldon v. Commissioner, supra. The inquiry into whether there was a legitimate business purpose involves a subjective analysis of the taxpayer's intent. See ACM Partnership v. Commissioner, supra at 247; Kirchman v. Commissioner, supra at 1492. We will begin with an examination of the economic substance of petitioner's 1993 COLI plan. In doing so, we focus on the COLI transaction in its entirety rather than any single step. See Kirchman v. Commissioner, supra at 1493-1494. Petitioner's 1993 purchase of COLI on the lives of approximately 36,000 of its employees was done pursuant to an 37 In certain situations courts have held that a transaction that lacks economic substance, other than the production of a tax benefit, is a substantive sham regardless of the motive of the taxpayer. See Knetsch v. United States, supra at 365; Dewees v. Commissioner, 870 F.2d 21, 35 (1st Cir. 1989); Kirchman v. Commissioner, supra at 1492. - 44 -overall plan that projected costs and benefits for each year over a 60-year period. See appendixes A and B. Petitioner also recognized that circumstances might well change during that period that would cause it to modify or terminate the plan. In fact, the COLI plan was impacted by legislation in 1996, and the COLI policies were terminated in 1997. However, for the first 2 years, the COLI plan was followed and it produced results that were consistent with plan projections.38 We will, therefore, examine the economic substance of the COLI transactions by analyzing the projections that reflect the plan. Shortly after having been approached by WJ/Coventry regarding proposals for COLI to be purchased from AIG, petitioner decided that it was interested in what was described as a "zero-cash strategy". This strategy was based on an elaborate plan involving the purchase of life insurance on the lives of over 36,000 of petitioner's then current employees. The plan was complex and depended upon relationships between many factors, including number of lives insured, premium levels, policy expenses, rates of interest to be charged and credited, policy loans, cash surrender values, withdrawals from cash surrender 38 The instant case involves deductions for accrued interest and fees in the first plan year. The first year of the COLI insurance began on Mar. 1, 1993, and ended on Feb. 28, 1994. The deductions in issue were based on an allocation of the interest and fees that had accrued during petitioner's taxable year ended June 30, 1993. - 45 -values, and death benefits. Petitioner was to be the owner and beneficiary of the policies. Detailed projections were prepared to demonstrate the financial impact of the plan. The projections assumed a high rate of interest (11.06 percent) would be charged to petitioner on its policy loans. This would be countered by a high rate of interest to be credited to petitioner on the portion of the gross cash surrender value that petitioner had borrowed against. The crediting rate was 40 basis points below the rate charged to petitioner on its policy loans (10.66 percent). The rate to be credited on the unborrowed portion of the gross cash surrender value was 4 percent. Policy loans by petitioner would be used to pay most of the premiums and interest with the result that petitioner's net equity in the policies would remain relatively small. Death benefits would be applied to reduce outstanding policy loans. The profit and loss statements in the projections illustrate the pretax effect and the after-tax effect that the COLI plan would have on petitioner. The difference between pretax and after-tax effects was based on the income tax savings that would result from deducting policy loan interest and administrative fees. Policy loan interest was clearly the dominant element. All the various projections prepared before the actual purchase of the policies in June 1993 show that the pretax effect on - 46 -petitioner for each policy year was a loss and that the after-tax effect was a significant profit. The projections submitted to petitioner on June 4, 1993, were prepared just before petitioner's purchase of the COLI policies in June 1993. These projections are attached as appendix B. We shall use figures from the projections in appendix B to illustrate the COLI plan's lack of economic substance. The elements of the COLI plan and their projected impact on petitioner at the completion of the first policy year were as follows. Petitioner would make a premium payment of $108,573,000 and simultaneously borrow $101,328,000 against the policy. This required petitioner to pay the balance of $7,245,000 to AIG to satisfy the premium. At the end of the policy year, interest accrued on petitioner's policy loans would be $11,191,000, and petitioner would also have incurred administrative fees of $290,000. What benefit was petitioner to get for these costs? At the end of the first policy year, the COLI policies would have net cash surrender value of $11,287,000. In addition, based on actuarial determinations, petitioner expected death benefits from the COLI policies in the first year to be $3,250,000.39 Based on the combination of these first-year costs and benefits, the net 39 Under the terms of the policies, death benefits from a policy would first be used to reduce any outstanding loan. - 47 -effect of the COLI plan was a first-year loss40 of $4,188,000 computed as follows: Net premium payment $7,245,000 Interest on policy loan 11,191,000 Administrative fees 290,000 18,726,000 Less: Net cash surrender value 11,287,000 Death benefits 3,250,000 14,537,000 1 Loss 4,189,000 1 The June 1993 projection shows $4,188,000. This is apparently due to rounding or a math error. Following the same approach, the June 1993 projections show the COLI plan producing pretax losses in the next 2 policy years of $7,885,000 and $10,869,000, respectively. Thereafter, the pretax losses over the next 57 years range from $6,244,000 in the last year to $16,447,000 in year 2021. The total of pretax losses for the projected 60 years was $681,922,000. In each and every year, the combined yearly pretax benefits from the policies; i.e., the expected death benefits from the 36,000 policies plus the year-end net equity value of the policies, were substantially less than petitioner's cost of maintaining the policies. The next part of the June 1993 profit and loss projections illustrates the "tax effect" of the COLI plan. The profit and loss statement contained in the June 1993 projections shows 40 The projections refer to the loss as negative pretax earnings. - 48 -first-year income tax savings from the COLI plan of $4,480,000. This amount is composed of tax savings of $4,368,000 attributable to a deduction of accrued first-year interest on policy loans of $11,191,000 and tax savings of $113,000 attributable to a deduction of first-year administrative fees of $290,000. Based on this, the projected "after-tax earnings effect" for the first policy year was $292,000.41 Similar projections for each of the following 59 years show that while the "pretax earnings effect" of the plan resulted in losses, the "after-tax earnings effect" continued to be positive in each year reaching its peak in the year 2008 when the "after-tax earnings effect" would be $63,965,000. This amount was arrived at by subtracting the pretax loss of $9,182,000 from projected income tax savings of $73,146,000.42 The projected income tax savings of $73,146,000 were attributable to tax deductions for interest of $187,279,000 and administrative fees of $276,000. The June 1993 projections indicate that had the 1993 COLI plan remained in effect through the year 2052, petitioner's total pretax loss over 60 years would have been $681,922,000 but that the total tax saved because of policy loan interest and fee deductions would have exceeded $3 41 The above figures were taken from the June 1993 projections reflected in appendix B. The totals vary by $1,000, apparently due to rounding off the last three digits. 42 See supra note 41. - 49 -billion, resulting in a total "after-tax earnings effect" of more than $2 billion. The June 1993 projections contain a cash-flow analysis for each policy year from 1993 to 2052. The structure of the zero-cash strategy was intended to produce a positive after-tax cash-flow for each policy year. Thus for the first year, the plan was to produce a positive cash-flow of $196,000 after factoring in tax savings from deducting policy loan interest and fees.43 Without the savings from these deductions, there would have been a negative cash-flow of over $4 million. The projections show increasing positive after-tax cash-flows for each of the following 59 years. Projected after-tax cumulative cash-flow for the entire 60-year period was more than $2 billion. Cumulative net equity at the end of each year varied, rising in some years and falling in others but, because of the policy loans and withdrawals, remained relatively small in relationship to the numbers in the overall plan. For example, cumulative net equity 43 In addition, the plan would result in petitioner’s having a cumulative net equity in the COLI policies at the end of the first policy year of $96,000. Cumulative net equity was the gross surrender value of the policies minus outstanding policy loans and accrued policy loan interest. Gross cash surrender value of $112,471,000 minus the sum of the outstanding loan of $101,184,000 and accrued loan interest of $11,191,000 equals $96,000. See appendix B, Balance Sheet Summary. The combination of cumulative net equity and positive cash-flow equals the projected positive after-tax earnings effect of $292,000. $96,000 plus $196,000 equals $292,000. - 50 -after 15 years was projected to be $498,000, whereas cumulative positive cash-flow was projected to be $289,263,000. See appendix B, Cash Flow. Without the tax savings from tax deductions for policy loan interest and fees, there would have been a substantial negative cash-flow in each year, and the costs of maintaining the COLI plan would have greatly exceeded benefits. 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. Thus, the predictable cost of maintaining life insurance might be greater than predictable death benefits and still be justified by the financial protection that insurance provides against the financial consequences of the unexpected death of the insured. But as we discuss later, petitioner had no such reason or purpose for engaging in the 1993 COLI program. Petitioner suggests that the policies could conceivably produce tax-independent benefits if some catastrophe were to occur that would produce large, unexpected death benefits. We are convinced that this was so improbable as to be unrealistic and therefore had no economic significance. Indeed, petitioner makes no pretense that it purchased these policies in anticipation of, or to protect itself against, a catastrophic event. The policies were on the lives of 36,000 individual employees of various ages who lived in diverse - 51 -locations. The insured employees' lives were to remain insured even after their employment was terminated. The anticipated mortality of this large group was actuarially determined, and both AIG and petitioner engaged in the COLI transactions based on these actuarial expectations. While there would obviously be some variation in the actual mortality of the insured population, such variations were not expected to significantly affect the plan. And as explained later, the function of the claims stabilization reserve was to ameliorate fluctuations in actual mortality experience. Economic substance depends on whether, from an objective standpoint, the transaction was likely to produce economic benefits aside from tax deductions. See Kirchman v. Commissioner, 862 F.2d at 1492; Bail Bonds by Marvin Nelson, Inc. v. Commissioner, 820 F.2d 1543, 1549 (9th Cir. 1987), affg. T.C. Memo. 1986-23. Viewing the COLI plan as a whole, we find that the only function of the plan was to produce tax deductions in order to reduce petitioner's income tax liabilities. 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. Consequently, the COLI transactions lacked economic substance apart from producing tax deductions. - 52 - In determining whether a transaction should be respected for tax purposes, we also look to whether the taxpayer had a business purpose for engaging in the transaction other than tax avoidance. See Frank Lyon Co. v. United States, 435 U.S. 561, 583-584 (1978); Kirchman v. Commissioner, supra at 1492; Bail Bonds by Marvin Nelson, Inc. v. Commissioner, supra at 1549. Petitioner argues that it had an economic objective and valid business purposes for entering into the COLI transaction other than tax avoidance. Petitioner alleges that before entering into the 1993 COLI transaction, it had become concerned with increasing costs associated with its Winn-Flex program and that it decided to implement the COLI program as a mechanism for obtaining funds to pay such costs. Before entering into the COLI transaction, there were numerous versions of profit and loss and cash-flow projections, which were consistently formatted so that petitioner could compare the pretax earnings effect to the post-tax earnings effect. Petitioner requested multiple versions of the projections at various estimated combined Federal and State marginal tax rates in order to see what effect a change in rates would have on the proposed COLI transaction. On the other hand, petitioner produced no contemporaneously prepared documents indicating that it purchased the 1993 COLI policies in order to provide a source for funding its Winn-Flex obligations. Unlike - 53 -the policies used to fund petitioner's obligations under its Management Security Program, the individual 1993 COLI policies were not tailored to fund benefits due the insured employees under Winn-Flex. Indeed, the policies were to remain in effect after the individual employees left petitioner's employ. In planning for and setting up the COLI plan, petitioner's financial vice president and principal financial officer, Mr. McCook, never told the individuals at WJ/Coventry, who were planing the COLI transactions, about any purpose or objective to use the COLI plan to fund benefits under Winn-Flex. On brief, petitioner argues that death benefits and policy loans and withdrawals from the net cash value of COLI policies could be used to help fund Winn-Flex. However, the projections, which embody petitioner's broad-based COLI plan, show that anticipated death benefits and net cash values were going to be exhausted in order to satisfy petitioner's premiums and policy loan interest obligations. According to petitioner's COLI plan, there would be no death benefits and cash value left over to provide the necessary funding for Winn-Flex. Indeed, the COLI plan anticipated that after using available death benefits, policy loan proceeds, and withdrawals, petitioner would still be required to make annual cash payments in order to satisfy its annual premium and policy loan interest obligations. We do not - 54 -believe that petitioner purchased the COLI policies to fund Winn-Flex. In his testimony, Mr. McCook made it clear that his focus was on the bottom line, after-tax earnings impact, of the COLI plan and the resulting positive cash-flow that the tax deductions were expected to generate. Referring to the January 27, 1993, projections of profit and loss prepared by Coventry (appendix A), Mr. McCook testified that he expected that by the 15th year the annual financial benefit of the COLI transaction would offset the annual costs of petitioner's Winn-Flex obligations. According to the January 27, 1993, projection of profit and loss (appendix A), there was a pretax loss in each year of the 60 years in the projection. The pretax loss for the 15th year (2007) was $14,178,000, and the cumulative pretax loss for the first 15 years was $148,483,000. The January 27, 1993, projection of profit and loss showed a profit for the year 2007 only after considering the tax savings from the policy loan interest and fee deductions. The projected after-tax profit from the COLI plan for 2007 was $64,479,000. When Mr. McCook identified the amount he believed would be available to fund the annual costs of Winn-Flex, he referred to the $64,479,000 amount of projected after-tax earnings from the COLI program for the year 2007. This amount was produced by loan interest and fee deductions. A tax savings generated by the COLI plan was the only reason the plan - 55 -produced positive earnings and cash-flow. Indeed, petitioner's internal records show that petitioner viewed the 1993 COLI plan as an "Expense Reduction Opportunity", that would produce estimated savings of $300 million.44 The only "expense" that was reduced by the COLI plan was petitioner's income tax liability. Even if we were to accept Mr. McCook's testimony that he intended to use tax savings to fund Winn-Flex, that would not cause the COLI plan to have economic substance.45 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 have been used as a general source of funds for petitioner's Winn-Flex obligations (or any other business purpose) does not alter the fact that the COLI plan 44 When Mr. McCook was asked how the $300 million was derived, he testified that he thought that it was the total of the "After-tax Savings" figures listed in the Jan. 27, 1993, projections. These Jan. 27, 1993, projections are contained in appendix A, Profit and Loss Statement. The after-tax earnings referred to by Mr. McCook are in column J. The total after-tax earnings for the policy years 1993 through 2007 are slightly more than $300 million. The total projected after-tax earnings for the years 1993 through 2052 are more than $2 billion. 45 We note that none of these tax savings were earmarked for funding Winn-Flex. They were simply projected to reduce petitioner's tax liabilities and thereby increase petitioner's after-tax profits by more than $2 billion over 60 years. - 56 -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. Petitioner also argues that the purchase of the COLI policies permitted it to increase group life benefits offered to Winn-Flex participants. It is true that petitioner offered an additional $5,000 in life insurance benefits to employees who agreed to allow petitioner to purchase COLI policies on their lives. However, this was done at the suggestion of Coventry in order to obtain the employees' consent to have their lives insured. There was no relationship between death benefits under the COLI policies and the relatively small $5,000 employee death benefit. All policies bore a $3,000 annual premium, and death benefits under the policies were based on that premium amount and the age of the employee. Also, petitioner had a high turnover among its employees, and the $5,000 death benefit expired when the insured's employment with petitioner ended. As a result, Coventry advised petitioner that the additional $5,000 in coverage could be provided at an insignificant cost. Based on the record, we do not believe that the purpose of the COLI plan was to fund employee benefits. Petitioner's COLI plan required a relatively small amount of cash investment by petitioner and charged a high rate of interest - 57 -on petitioner's policy loans based on the assumption that petitioner's "appetite for interest deductions remains large". The projections showed that the COLI plan would generate positive cash-flows and earnings only because of the tax benefit associated with the interest and fee deductions. Tax considerations permeated the planning stages of petitioner's COLI. When the broad-based COLI plan was first explained to him, Mr. McCook recognized that it was a tax shelter. Mr. McCook's primary concern was to achieve a positive cash-flow. The only way a positive cash-flow could be achieved was through the deduction of interest on policy loans. This is why petitioner concentrated on its ability to deduct loan interest and the availability of "exit strategies" in the event new legal restrictions on deductions were enacted or petitioner's "appetite" for interest deductions diminished. Following the enactment of tax law changes in August 1996, which greatly restricted employers' deductions for interest on loans from company-owned life insurance policies on the lives of employees, petitioner terminated its COLI program. See Health Insurance Portability and Accountability Act of 1996, Pub. L. 104-191, sec. 501, 110 Stat. 2090. The 1996 change in the tax law caused petitioner's COLI program to become a financial burden because it specifically prohibited the deduction of policy loan interest under petitioner's plan. After the 1996 tax law change, - 58 -none of petitioner's purported business purposes affected petitioner's decision to terminate the COLI program. Petitioner cites Campbell v. Cen-Tex, Inc., 377 F.2d 688 (5th Cir. 1967), as controlling precedent in this case.46 Petitioner's reliance on this case is misplaced. Cen-Tex was a family-owned corporation that had entered into deferred compensation arrangements, which obligated it to provide payments to the surviving spouse or lineal descendants of employee stockholders and to purchase and redeem stock of deceased stockholders. Cen-Tex decided to meet these obligations by purchasing insurance on the lives of the employee stockholders. Cen-Tex paid the first annual premium on each policy and prepaid the next four annual premiums, discounted at 3 percent, and then borrowed against the value on each of the policies at a 4-percent rate. See id. at 689. The court allowed deductions for interest on the policy loans. Cen-Tex, Inc. is clearly distinguishable from petitioner's case. The parties in Cen-Tex, Inc. stipulated that the insurance policies at issue were procured to assist in meeting the obligations of the taxpayer under its deferred compensation plan and its obligations under the stock option and redemption 46 Petitioner's case is appealable to the Court of Appeals for the Eleventh Circuit. Decisions of the Court of Appeals for the Fifth Circuit that were handed down prior to Sept. 30, 1981, are generally binding as precedent in the Eleventh Circuit. Bonner v. City of Pritchard, 661 F.2d 1206 (11th Cir. 1981). - 59 -agreement, as well as for the general objective of having insurance on its key employees and stockholders. Based on this concession, the court found there was a bona fide nontax business purpose and economic objective to be served by the insurance. See id. The court also found that the transaction produced benefits other than tax benefits. The court concluded that "The policies purchased provided for a beneficial interest. The transaction was not without economic value, economic significance, economic substance, or commercial substance." Campbell v. Cen-Tex, Inc., supra at 692 (fn. refs. omitted). In contrast to Campbell v. Cen-Tex, Inc., supra, we have found that no nontax purpose was served by the COLI transactions. The projections for the COLI policies contemplated a substantial pretax loss in each year, even after considering the projected death benefits and net cash surrender value of the policies. Only by deducting the policy loan interest and fees and reducing its income tax could petitioner anticipate any benefit from its COLI transactions. Without the tax benefits of the policy loan interest and fee deductions being generated by the COLI plan, petitioner's plan would have generated a predictable negative cash-flow and pretax loss in each of the 60 years projected. This predictable result precludes any economic value, economic - 60 -significance, economic substance, or commercial substance other than the tax benefit. Based upon all the aforementioned considerations, we find that petitioner purchased the COLI policies in 1993 pursuant to a plan the only function of which was to generate interest and fee deductions in order to offset income from other sources and thereby significantly reduce its income tax liability. We hold that petitioner's 1993 broad-based COLI program lacked substance and was a sham. Petitioner argues that lack of economic substance does not warrant disallowing the interest deduction in question because deductions for interest on life insurance policy loans were condoned by Congress as indicated by the safe harbor test of section 264 and its legislative history. Petitioner argues that the legislative history shows that Congress clearly accepted the deductibility of interest on corporate-owned life insurance products that satisfy the safe harbor tests of section 264. Petitioner maintains that because its COLI policies were life insurance contracts within the meaning of section 7702 and its pattern of borrowing from the policies satisfied the "four-of-seven test" of section 264(c)(1), its loan interest is deductible under section 163. Petitioner also argues that because Congress, through legislation in 1996, further extended its denial of deductions associated with interest payments on COLI policy - 61 -loans, petitioner was not barred from taking such deductions prior to 1996. Section 264(a)(3) generally provides that no deduction is allowed for amounts paid or accrued on indebtedness incurred to purchase a life insurance contract if such debt was incurred pursuant to a plan of purchase which contemplates the systematic borrowing of increases in the cash value of the insurance contract. Section 264(c)(1) provides an exception to the general rule under section 264(a)(3). Section 264(c)(1) provides that if no part of any four annual premiums due in the first 7-year period of an insurance contract is financed by means of indebtedness, then the general rule of section 264(a)(3) will not apply.47 47 In pertinent part, sec. 264(a) provides: SEC. 264(a) General Rule.--No deduction shall be allowed for-- • * * * * * * (3) Except as provided in subsection (c), any amount paid or accrued on indebtedness incurred or continued to purchase or carry a life insurance, endowment, or annuity contract (other than a single premium contract or a contract treated as a single premium contract) pursuant to a plan of purchase which contemplates the systematic direct or indirect borrowing of part or all of the increases in the cash value of such contract (either from the insurer or otherwise). - 62 -The parties refer to this exception as the "four-of-seven test". The parties agree that petitioner's COLI policies meet the requirements of the four-of-seven test.48 The parties disagree • * * * * * * (c) Exceptions.--Subsection (a)(3) shall not apply to any amount paid or accrued by a person during a taxable year on indebtedness incurred or continued as part of a plan referred to in subsection (a)(3)-- (1) if no part of 4 of the annual premiums due during the 7-year period (beginning with the date the first premium on the contract to which such plan relates was paid) is paid under such plan by means of indebtedness, (2) if the total of the amounts paid or accrued by such person during such taxable year for which (without regard to this paragraph) no deduction would be allowable by reason of subsection (a)(3) does not exceed $100, (3) if such amount was paid or accrued on indebtedness incurred because of an unforeseen substantial loss of income or unforeseen substantial increase in his financial obligations, or (4) if such indebtedness was incurred in connection with his trade or business. For purposes of applying paragraph (1), 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. 48 The parties also agree that petitioner's COLI policies meet the definition of a life insurance contract for purposes of - 63 -as to whether satisfaction of the requirements of section 264(a) and (c) authorizes a deduction of the interest expenses arising out of a transaction that otherwise is without substance. An argument similar to petitioner's was made in Knetsch v. United States, 364 U.S. 361 (1960). In Knetsch, the Court found that the taxpayer's purchase of annuity contracts and simultaneous loans from an insurance company was a sham that did not give rise to deductible interest. Nevertheless, like petitioner, the taxpayer in Knetsch contended that by enacting section 264 as part of the 1954 Code, Congress "authorized" the interest deductions for transactions prior to the effective date of the 1954 Code. See id. at 367. Section 264(a)(2), as enacted in 1954, denied a deduction for amounts paid on indebtedness incurred to purchase or carry a single premium annuity contract, but only as to contracts purchased after March 1, 1954, the date of enactment. See id. From this the taxpayers reasoned that Congress intended to allow interest deductions for such transactions occurring prior to March 1, 1954, regardless of their substance. The Supreme Court disagreed, concluding that unless such meaning plainly appeared from the statute and its legislative history, the Court would not attribute such an intent to Congress, for "'To hold otherwise would be to exalt artifice above reality and to deprive the statutory provision in question sec. 7702. - 64 -of all serious purpose.'" Id. (quoting Gregory v. Helvering, 293 U.S. at 470). A taxpayer's right to a deduction for interest on an insurance policy loan is based on section 163, not section 264. Golsen v. Commissioner, 54 T.C. 742, 755 (1970), affd. 445 F.2d 985 (10th Cir. 1971). Section 264 does not confer the right to a deduction but simply denies, disallows, or prohibits deductions that might otherwise be allowable under some other provision. See id. at 756. Thus, while the parties agree that petitioner's COLI plan meets the "four-of-seven test" of section 264(c)(1) and would be excepted from the general disallowance rule of section 264(a)(3), section 264 does not confer a right upon petitioner to take the deduction that would not otherwise be allowable under section 163. Petitioner cites the Senate Finance Committee's report discussing the scope of section 264 prior to the 1964 amendment. The report states that "under present law, no interest deductions are denied where the taxpayer purchases an insurance contract with the intention of borrowing the maximum amount on the contract each year". S. Rept. 830, 88th Cong., 2d Sess. (1964), 1964-1 C.B. (Part 2) 505, 581. Based on this, petitioner argues that Congress did not view the Supreme Court's decision in Knetsch as foreclosing interest deductions based on the type of - 65 -sham transactions involved in this case. A similar argument was advanced in McLane v. Commissioner, 46 T.C. 140 (1966), affd. 377 F.2d 557 (9th Cir. 1967), where the taxpayers had engaged in a series of transactions similar to those in the instant case. In the Revenue Act of 1964, Pub. L. 88-272, sec. 215(a), 78 Stat. 55, Congress added subsection (a)(3) of section 264 to address problems associated with amounts paid or accrued on indebtedness incurred with respect to several types of insurance contracts pursuant to a plan of systematic borrowing. In McLane v. Commissioner, supra at 144-145, we considered the same passage from the Senate Finance Committee report that petitioner cites and stated: Based upon the foregoing, petitioner by a tour de force concludes that: (a) The 1958 transaction herein is the type of abuse meant to be curbed by subsection (a)(3), but only prospectively; (b) the legislative history expressly confirms his assertion that the deduction flowing from this abuse was allowable under prior law; and (c) the 'interest' involved herein is therefore deductible. We agree with petitioners that the 1958 transaction in form fell within the class of transactions at which subsection (a)(3) was aimed. But we do not agree with his assertion that the legislative history should be turned into an open-ended license applicable without regard to the substance of the transaction. Nor do we agree with the assertion that, if Knetsch and Pierce, were controlling with respect to post-1958 multiple-premium annuities, there would have been no need for further legislation in 1964. Knetsch and Pierce involved transactions without substance. Congress, in enacting section 264(a)(3), struck at transactions with substance. It is a reductio ad absurdum to reason, as petitioner does, that Congress simultaneously struck down a warm body and breathed - 66 - life into petitioner's cadaver. [Fn. ref. omitted.] Petitioner attempts to supplement its argument by citing additional legislative materials related to changes or proposed changes to section 264 in 1984, 1986, 1987, 1988, 1990, 1991, and 1996. We need not address each of the changes and proposals regarding interest deductions on life insurance policy loans. It is clear that Congress and the Treasury Department were aware of the problems associated with interest deductions on life insurance loans. However, 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. In Knetsch, the Supreme Court noted that nothing in the legislative history of section 264 suggests that Congress intended to protect sham transactions. Similarly, we find nothing in the more recent legislative history of section 264 suggesting that Congress intended to allow deductions arising from sham transactions that lacked economic substance and business purpose. The transactions associated with petitioner's COLI program lacked economic substance and business purpose (other than tax reduction). As a result, the interest on petitioner's COLI loans was not deductible interest on indebtedness within the meaning of - 67 -section 163. The same reasoning applies to the administrative fees associated with the COLI plan.49 They were incurred in connection with, and were an integral part of, a sham transaction and, as a result, are not deductible. See Karr v. Commissioner, 924 F.2d at 1022-1023; Kirchman v. Commissioner, 862 F.2d 1486; Lee v. Commissioner, 155 F.3d 584 (2d Cir. 1998). We, therefore, uphold respondent's disallowance of these deductions. Decision will be entered under Rule 155. 49 Respondent argues that the administrative fees should be disallowed pursuant to sec. 265. Because we have held that the administrative fees must be disallowed as the product of a sham, we have no need to consider disallowance under sec. 265. - 68 - Appendix A Scenario 1 - Constant Loan Interest Rate Profit and Loss Statement (dollars in thousands except earnings per share) Pre-Tax Effect Tax Effect (A) (B) (C) (C1) (D) (E) (F) (G) (H) (I) (J) (K) Annual Accrued Deductible Pre-Tax Policy Admin. Net CSV Loan Loan Earnings Loan Fee Tax After-Tax After-Tax Annual Increase/ Interest Interest Death Admin. Effect Tax Tax Effect Earnings Earnings Year (Premium)* (Decrease) (Payment) (Payment) Benefits Fee A+B+C+D+E Credit Credit G+H Effect Per Share**[1] 1993 (114,000) 119,586 (11,902) (11,902) 2,016 (304) (4,605) 4,524 116 4,640 35 0.00 1994 (112,280) 126,513 (24,486) (24,486) 2,155 (304) (8,403) 9,308 115 9,423 1,021 0.01 1995 (99,121) 122,492 (36,661) (36,661) 2,312 (304) (11,282) 13,936 115 14,052 2,770 0.04 1996 23,952 (29) (36,633) (36,633) 2,614 (303) (10,399) 13,926 115 14,041 3,642 0.05 1997 24,182 (30) (36,602) (36,602) 2,756 (303) (9,997) 13,915 115 14,030 4,033 0.05 1998 24,411 (32) (36,570) (36,570) 2,934 (303) (9,559) 13,903 115 14,018 4,458 0.06 1999 24,340 (35) (36,535) (36,535) 3,152 (303) (9,381) 13,890 115 14,005 4,624 0.06 2000 (113,380) 152,293 (51,654) (51,654) 3,287 (302) (9,756) 19,639 115 19,753 9,997 0.13 2001 (113,264) 168,363 (68,351) (68,351) 3,595 (302) (9,959) 25,988 115 26,103 16,143 0.21 2002 (113,137) 185,941 (86,771) (86,771) 3,959 (302) (10,310) 32,994 115 33,108 22,799 0.30 2003 (112,997) 205,253 (107,079) (107,079) 4,399 (301) (10,725) 40,718 115 40,833 30,108 0.39 2004 (112,841) 226,309 (129,436) (129,436) 4,911 (301) (11,358) 49,224 114 49,338 37,980 0.50 2005 (112,667) 249,264 (154,019) (154,019) 5,508 (300) (12,215) 58,578 114 58,692 46,477 O.61 2006 (112,472) 274,461 (181,034) (181,034) 6,194 (300) (13,151) 68,860 114 68,974 55,822 0.73 2007 (112,253) 302,091 (210,699) (206,693) 6,983 (299) (14,178) 78,543 114 78,657 64,479 0.84 2008 0 213,785 (231,471) (206,209) 7,851 (299) (10,134) 78,359 113 78,473 68,339 0.89 2009 0 234,409 (254,176) (205,666) 8,796 (298) (11,269) 78,153 113 78,266 66,998 0.87 2010 0 257,067 (278,992) (205,060) 9,803 (297) (12,420) 77,923 113 78,036 65,616 0.86 2011 0 281,888 (306,108) (204,387) 10,864 (296) (13,653) 77,667 113 77,780 64,127 0.84 2012 0 309,092 (335,731) (203,644) 11,961 (295) (14,973) 77,385 112 77,497 62,524 0.82 2013 0 338,897 (368,084) (202,828) 13,091 (294) (16,390) 77,075 112 77,186 60,796 0.79 2014 0 371,620 (403,420) (201,938) 14,253 (293) (17,839) 76,736 111 76,848 59,009 2015 0 408,917 (442,145) (200,971) 15,448 (291) (18,071) 76,369 111 76,480 58,409 2016 0 449,906 (484,575) (199,926) 16,673 (290) (18,285) 75,972 110 76,082 57,797 2017 0 494,902 (531,044) (198,799) 17,938 (288) (18,492) 75,544 110 75,653 57,161 0.75 2018 0 544,368 (581,916) (197,588) 19,289 (287) (18,546) 75,084 109 75,192 56,647 0.74 2019 0 598,625 (637,563) (196,285) 20,777 (285) (18,446) 74,588 108 74,696 56,250 0.73 2020 0 657,819 (698,339) (194,878) 22,461 (283) (18,342) 74,054 108 74,161 55,819 0.73 2021 0 722,246 (764,590) (193,354) 24,397 (281) (18,228) 73,474 107 73,581 55,353 0.72 2022 0 792,208 (836,648) (191,694) 26,615 (279) (18,103) 72,844 106 72,950 54,846 0.72 *Total annual premium less annual withdrawal. **Based on 76.6 million shares outstanding. All figures are estimates. Actual results will depend upon mortality, interest rates and dividends. __________________________________ 1 Blank space indicates that there was no legible figure in underlying exhibit. - 69 - Scenario 1 - Constant Loan Interest Rate Profit and Loss Statement (dollars in thousands except earnings per share) Pre-Tax Effect Tax Effect (A) (B) (C) (C1) (D) (E) (F) (G) (H) (I) (J) (K) Annual Accrued Deductible Pre-Tax Policy Admin. Net CSV Loan Loan Earnings Loan Fee Tax After-Tax After-Tax Annual Increase/ Interest Interest Death Admin. Effect Tax Tax Effect Earnings Earnings Year (Premium)* (Decrease) (Payment) (Payment) Benefits Fee A+B+C+D+E Credit Credit G+H Effect Per Share**[1] 2023 0 867,928 (914,804) (189,879) 29,184 (276) (17,968) 72,154 105 72,259 54,291 0.71 2024 0 949,613 (999,303) (187,885) 32,146 (273) (17,817) 71,396 104 71,500 53,683 0.70 2025 0 1,037,393 (1,090,314) (185,684) 35,546 (270) (17,645) 70,560 103 70,662 53,017 0.69 2026 0 1,131,328 (1,187,917) (183,247) 39,410 (267) (17,446) 69,634 101 69,735 52,289 0.68 2027 0 1,231,384 (1,292,076) (180,544) 43,740 (263) (17,215) 68,607 100 68,707 51,492 0.67 2028 0 1,337,543 (1,402,677) (177,547) 48,446 (259) (16,946) 67,468 98 67,567 50,620 0.66 2029 0 1,449,715 (1,519,519) (174,238) 53,415 (255) (16,643) 66,210 97 66,307 49,664 0.65 2030 0 1,567,742 (1,642,320) (170,601) 58,634 (250) (16,193) 64,829 95 64,923 48,730 0.64 2031 0 1,691,227 (1,770,698) (166,634) 64,628 (244) (15,086) 63,321 93 63,414 48,327 0.63 2032 0 1,819,004 (1,904,072) (162,337) 68,762 (238) (16,545) 61,688 90 61,778 45,233 0.59 2033 0 1,953,020 (2,041,845) (157,710) 73,819 (232) (15,238) 59,930 88 60,018 44,780 0.58 2034 0 2,089,112 (2,182,737) (152,749) 79,081 (225) (14,769) 58,044 85 58,130 43,361 0.57 2035 0 2,226,401 (2,325,080) (147,440) 84,576 (217) (14,320) 56,027 83 56,110 41,789 0.55 2036 0 2,362,737 (2,466,685) (141,768) 90,329 (209) (13,828) 53,872 80 53,952 40,124 0.52 2037 0 2,495,443 (2,604,815) (135,720) 96,292 (201) (13,281) 51,574 76 51,650 38,369 0.50 2038 0 2,621,661 (2,736,378) (129,288) 102,236 (192) (12,674) 49,130 73 49,203 36,529 O.48 2039 0 2,738,347 (2,858,051) (122,486) 107,873 (182) (12,013) 46,545 69 46,614 34,601 0.45 2040 0 2,842,259 (2,966,346) (115,340) 112,947 (172) (11,312) 43,829 65 43,895 32,582 0.43 2041 0 2,930,135 (3,057,756) (107,898) 117,200 (162) (10,583) 41,001 61 41,063 30,479 0.40 2042 0 2,998,802 (3,128,893) (100,220) 122,567 (151) (7,675) 38,084 57 38,141 30,466 0.40 2043 0 3,033,490 (3,175,595) (92,381) 136,378 (139) (5,867) 35,105 53 35,158 29,291 2044 0 3,035,085 (3,194,629) (84,469) 151,298 (128) (8,374) 32,098 49 32,147 23,772 2045 0 3,010,969 (3,184,372) (76,577) 165,738 (116) (7,782) 29,099 44 29,143 21,361 2046 0 2,957,486 (3,143,787) (68,803) 179,192 (105) (7,214) 26,145 40 26,185 18,971 0.25 2047 0 2,873,589 (3,071,784) (61,234) 191,639 (94) (6,650) 23,269 36 23,305 16,655 0.22 2048 0 2,760,378 (2,968,747) (53,949) 202,343 (83) (6,110) 20,501 32 20,532 14,423 0.19 2049 0 2,619,159 (2,835,702) (47,020) 211,032 (73) (5,583) 17,868 28 17,895 12,313 0.16 2050 0 2,451,775 (2,674,269) (40,506) 217,487 (63) (5,070) 15,392 24 15,416 10,347 0.14 2051 0 2,261,733 (2,487,427) (34,457) 221,162 (54) (4,586) 13,094 21 13,114 8,529 0.11 2052 0 2,102,864 (2,284,848) (28,917) 177,288 (46) (4,742) 10,988 17 11,006 6,264 0.08 *Total annual premium less annual withdrawal. **Based on 76.6 million shares outstanding. All figures are estimates. Actual results will depend upon mortality, interest rates and dividends. _________________________________ 1 Blank space indicates that there was no legible figure in underlying exhibit. - 70 - Scenario 1 - Constant Loan Interest Rate Cash Flow Detail (dollars in thousands) Corporate Cash Outflow Corporate Cash Inflow Net Cash Flow Surplus (A) (B) (C) (D) (E) (F) (G) (H) (I) (J) (K) Cumulative Cash After-Tax Policy Loan Tax-Free Net Cumulative Flow Cumulative Loan Admin. Tax Policy Policy Death Cash Cash at 4.35% Net Year Premium Interest Fee Savings Loan Withdrawal Benefits Clow Flow Pre-Tax* Equity[1] 1993 (114,000) 0 (188) 4,524 107,684 0 2,016 35 35 (53) 78 1994 (113,929) (11,902) (188) 9,308 113,929 1,649 2,155 1,021 1,056 838 178 1995 (113,852) (24,486) (188) 13,936 110,318 14,731 2,312 2,770 3,826 3,485 237 1996 (113,771) (36,661) (188) 13,926 0 137,722 2,614 3,642 7,468 7,095 337 1997 (113,681) (36,633) (188) 13,915 0 137,864 2,756 4,033 11,501 11,201 331 1998 (113,588) (36,602) (188) 13,903 0 137,999 2,934 4,458 15,959 15,854 175 1999 (113,488) (36,570) (188) 13,890 0 137,828 3,152 4,624 20,583 20,799 133 2000 (113,380) (36,535) (187) 19,639 137,175 0 3,287 9,997 30,580 31,316 242 2001 (113,264) (51,654) (187) 25,988 151,666 0 3,595 16,143 46,723 48,339 266 2002 (113,137) (68,351) (187) 32,994 167,521 0 3,959 22,799 69,522 72,556 295 2003 (112,997) (86,771) (187) 40,718 184,946 0 4,399 30,108 99,630 104,822 327 2004 (112,841) (107,079) (187) 49,224 203,952 0 4,911 37,980 137,610 145,922 367 2005 (112,667) (129,436) (186) 58,578 224,681 0 5,508 46,477 184,087 196,723 414 2006 (112,472) (154,019) (186) 68,860 247,446 0 6,194 55,822 239,910 258,345 466 2007 (112,253) (181,034) (186) 78,543 272,425 0 6,983 64,479 304,388 330,377 524 2008 0 (210,699) (185) 78,359 193,013 0 7,851 68,339 372,728 408,310 519 2009 0 (231,471) (185) 78,153 211,704 0 8,796 66,998 439,725 486,960 575 2010 0 (254,176) (184) 77,923 232,250 0 9,803 65,616 505,341 566,303 635 2011 0 (278,992) (184) 77,667 254,772 0 10,864 64,127 569,468 646,249 700 2012 0 (306,108) (183) 77,385 279,469 0 11,961 62,524 631,992 726,695 769 2013 0 (335,731) (182) 77,075 306,543 0 13,091 60,796 692,788 807,528 2014 0 (368,084) (182) 76,736 336,285 0 14,253 59,009 751,797 888,697 2015 0 (403,420) (181) 76,369 370,192 0 15,448 58,409 810,205 971,429 2016 0 (442,145) (180) 75,972 407,477 0 16,673 57,797 868,002 1,055,756 920 2017 0 (484,575) (179) 75,544 448,433 0 17,938 57,161 925,163 1,141,696 950 2018 0 (531,044) (178) 75,084 493,496 0 19,289 56,647 981,810 1,229,415 978 2019 0 (581,916) (177) 74,588 542,978 0 20,777 56,250 1,038,060 1,319,082 1,005 2020 0 (637,563) (175) 74,054 597,043 0 22,461 55,819 1,093,880 1,410,713 1,035 2021 0 (698,339) (174) 73,474 655,995 0 24,397 55,353 1,149,233 1,504,320 1,069 2022 0 (764,590) (173) 72,844 720,151 0 26,615 54,846 1,204,080 1,599,912 1,107 *Assumes deaths occur midyear. All figures are estimates. Actual results will depend upon mortality, interest rates and dividends. _______________________________ 1 Blank space indicates that there was no legible figure in underlying exhibit. - 71 - Scenario 1 - Constant Loan Interest Rate Cash Flow Detail (dollars in thousands) Corporate Cash Outflow Corporate Cash Inflow Net Cash Flow Surplus (A) (B) (C) (D) (E) (F) (G) (H) (I) (J) (K) Cumulative Cash After-Tax Policy Loan Tax-Free Net Cumulative Flow Cumulative Loan Admin. Tax Policy Policy Death Cash Cash at 4.35% Net Year Premium Interest Fee Savings Loan Withdrawal Benefits Flow Flow Pre-Tax* Equity[1] 2023 0 (836,648) (171) 72,154 789,772 0 29,184 54,291 1,258,371 1,697,489 1,151 2024 0 (914,804) (169) 71,396 865,114 0 32,146 53,683 1,312,054 1,797,047 1,201 2025 0 (999,303) (168) 70,560 946,382 0 35,546 53,017 1,365,071 1,898,574 1,258 2026 0 (1,090,314) (166) 69,634 1,033,725 0 39,410 52,289 1,417,360 2,002,054 1,322 2027 0 (1,187,917) (163) 68,607 1,127,225 0 43,740 51,492 1,468,851 2,107,465 1,393 2028 0 (1,292,076) (161) 67,468 1,226,943 0 48,446 50,620 1,519,472 2,214,778 1,469 2029 0 (1,402,677) (158) 66,210 1,332,873 0 53,415 49,664 1,569,135 2,323,956 1,548 2030 0 (1,519,519) (155) 64,829 1,444,941 0 58,634 48,730 1,617,866 2,435,071 1,508 2031 0 (1,642,320) (151) 63,321 1,562,850 0 64,628 48,327 1,666,193 2,548,711 685 2032 0 (1,770,698) (148) 61,688 1,685,629 0 68,762 45,233 1,711,426 2,662,208 1,729 2033 0 (1,904,072) (144) 59,930 1,815,248 0 73,819 44,780 1,756,207 2,778,260 1,858 2034 0 (2,041,845) (139) 58,044 1,948,220 0 79,081 43,361 1,799,568 2,895,941 1,933 2035 0 (2,182,737) (135) 56,027 2,084,057 0 84,576 41,789 1,841,357 3,015,138 2,010 2036 0 (2,325,080) (130) 53,872 2,221,133 0 90,329 40,124 1,881,481 3,135,794 2,089 2037 0 (2,466,685) (125) 51,574 2,357,313 0 96,292 38,369 1,919,850 3,257,856 2,168 2038 0 (2,604,815) (119) 49,130 2,490,097 0 102,236 36,529 1,956,379 3,381,277 2,243 2039 0 (2,736,378) (113) 46,545 2,616,674 0 107,873 34,601 1,990,980 3,506,008 2,309 2040 0 (2,858,051) (107) 43,829 2,733,964 0 112,947 32,582 2,023,562 3,632,001 2,362 2041 0 (2,966,346) (100) 41,001 2,838,725 0 117,200 30,479 2,054,041 3,759,218 2,397 2042 0 (3,057,756) (93) 38,084 2,927,665 0 122,567 30,466 2,084,507 3,889,822 169 2043 0 (3,128,893) (86) 35,105 2,986,788 0 136,378 29,291 2,113,799 4,022,599 2044 0 (3,175,595) (79) 32,098 3,016,051 0 151,298 23,772 2,137,571 4,153,131 2045 0 (3,194,629) (72) 29,099 3,021,226 0 165,738 21,361 2,158,932 4,284,556 2046 0 (3,184,372) (65) 26,145 2,998,071 0 179,192 18,971 2,177,903 4,416,931 (4,341) 2047 0 (3,143,787) (58) 23,269 2,945,592 0 191,639 16,655 2,194,558 4,550,372 (4,598) 2048 0 (3,071,784) (52) 20,501 2,863,415 0 202,343 14,423 2,208,980 4,685,016 (4,796) 2049 0 (2,968,747) (45) 17,868 2,752,205 0 211,032 12,313 2,221,293 4,821,045 (4,928) 2050 0 (2,835,702) (39) 15,392 2,613,208 0 217,487 10,347 2,231,639 4,958,675 (4,991) 2051 0 (2,674,269) (34) 13,094 2,448,575 0 221,162 8,529 2,240,168 5,098,134 (4,975) 2052 0 (2,487,427) (28) 10,988 2,305,443 0 177,288 6,264 2,246,432 5,239,658 (4,941) *Assumes deaths occur midyear. All figures are estimates. Actual results will depend upon mortality, interest rates and dividends. __________________________________ 1 Blank space indicates that there was no legible figure in underlying exhibit. - 72 - Scenario 1 - Constant Loan Interest Rate Balance Sheet Summary (dollars in thousands) (A) (B1) (B2) (B) (C) (D) (E) Gross Insurance Net Accrued Retained Annual Cash Cash Surrender Outstanding Cash Surrender Loan Earnings Impact on Year Amount Value (Loan) Value Interest Gain/(Loss) Earnings 1993 35 119,596 (107,616) 11,980 11,902 113 113 1994 1,056 246,061 (221,396) 24,665 24,486 1,234 1,121 1995 3,826 368,374 (331,476) 36,898 36,661 4,063 2,828 1996 7,468 368,186 (331,216) 36,970 36,633 7,805 3,743 1997 11,501 367,876 (330,943) 36,933 36,602 11,831 4,026 1998 15,959 367,397 (330,652) 36,745 36,570 16,134 4,303 1999 20,583 367,008 (330,339) 36,669 36,535 20,717 4,582 2000 30,580 518,930 (467,034) 51,896 51,654 30,822 10,106 2001 46,723 686,622 (618,005) 68,618 68,351 46,990 16,168 2002 69,522 871,619 (784,552) 87,066 86,771 69,817 22,827 2003 99,630 1,075,569 (968,163) 107,406 107,079 99,957 30,140 2004 137,610 1,300,113 (1,170,310) 129,803 129,436 137,977 38,019 2005 184,087 1,547,013 (1,392,580) 154,433 154,019 184,501 46,524 2006 239,910 1,818,331 (1,636,831) 181,499 181,034 240,375 55,875 2007 304,388 2,116,279 (1,905,056) 211,223 210,699 304,912 64,537 2008 372,728 2,324,858 (2,092,868) 231,990 231,471 373,247 68,334 2009 439,725 2,552,907 (2,298,156) 254,751 254,176 440,300 67,054 2010 505,341 2,802,162 (2,522,535) 279,628 278,992 505,977 65,676 2011 569,468 3,074,513 (2,767,705) 306,808 306,108 570,168 64,192 2012 631,992 3,372,040 (3,035,541) 336,500 335,731 632,761 62,593 2013 692,788 3,696,994 (3,328,067) 368,927 368,084 693,631 60,870 2014 751,797 4,051,835 (3,647,555) 404,280 403,420 752,657 59,026 2015 810,205 4,440,729 (3,997,694) 443,035 442,145 811,096 58,438 2016 868,002 4,866,820 (4,381,326) 485,495 484,575 868,922 57,826 2017 925,163 5,333,474 (4,801,480) 531,993 531,044 926,113 57,191 2018 981,810 5,844,342 (5,261,448) 582,894 581,916 982,788 56,675 2019 1,038,060 6,403,152 (5,764,585) 638,568 637,563 1,039,065 56,277 2020 1,093,880 7,013,474 (6,314,100) 699,374 698,339 1,094,915 55,849 2021 1,149,233 7,678,773 (6,913,114) 765,659 764,590 1,150,302 55,387 2022 1,204,080 8,402,381 (7,564,627) 837,754 836,648 1,205,186 54,885 All figures are estimates. Actual results will depend upon mortality, interest rates and dividends. - 73 - Scenario 1 - Constant Loan Interest Rate Balance Sheet Summary (dollars in thousands) (A) (B1) (B2) (B) (C) (D) (E) Gross Insurance Net Accrued Retained Annual Cash Cash Surrender Outstanding Cash Surrender Loan Earnings Impact on Year Amount Value (Loan) Value Interest Gain/(Loss) Earnings 2023 1,258,371 9,187,244 (8,271,289) 915,955 914,804 1,259,522 54,335 2024 1,312,054 10,035,794 (9,035,290) 1,000,504 999,303 1,313,255 53,733 2025 1,365,071 10,949,751 (9,858,179) 1,091,572 1,090,314 1,366,329 53,074 2026 1,417,360 11,929,899 (10,740,660) 1,189,239 1,187,917 1,418,682 52,353 2027 1,468,851 12,975,894 (11,682,425) 1,293,469 1,292,076 1,470,244 51,563 2028 1,519,472 14,086,579 (12,682,433) 1,404,146 1,402,677 1,520,941 50,696 2029 1,569,135 15,259,936 (13,738,869) 1,521,067 1,519,519 1,570,684 49,743 2030 1,617,866 16,493,017 (14,849,189) 1,643,828 1,642,320 1,619,373 48,690 2031 1,666,193 17,781,309 (16,009,926) 1,771,382 1,770,698 1,666,878 47,504 2032 1,711,426 19,121,648 (17,215,847) 1,905,801 1,904,072 1,713,155 46,277 2033 1,756,207 20,505,234 (18,461,532) 2,043,703 2,041,845 1,758 064 44,909 2034 1,799,568 21,920,085 (19,735,415) 2,184,670 2,182,737 1,801,501 43,437 2035 1,841,357 23,349,517 (21,022,427) 2,327,091 2,325,080 1,843,368 41,867 2036 1,881,481 24,771,530 (22,302,756) 2,468,774 2,466,685 1,883,570 40,202 2037 1,919,850 26,158,656 (23,551,673) 2,606,983 2,604,815 1,922,018 38,448 2038 1,956,379 27,479,838 (24,741,217) 2,738,622 2,736,378 1,958,623 36,604 2039 1,990,980 28,701,693 (25,841,333) 2,860,360 2,858,051 1,993,289 34,667 2040 2,023,562 29,789,198 (26,820,490) 2,968,708 2,966,346 2,025,924 32,635 2041 2,054,041 30,707,139 (27,646,985) 3,060,154 3,057,756 2,056,439 30,514 2042 2,084,507 31,419,234 (28,290,172) 3,129,061 3,128,893 2,084,676 28,238 2043 2,113,799 31,884,697 (28,712,437) 3,172,260 3,175,595 2,110,463 25,787 2044 2,137,571 32,075,465 (28,884,537) 3,190,928 3,194,629 2,133,870 23,406 2045 2,158,932 31,972,124 (28,791,793) 3,180,331 3,184,372 2,154,891 21,022 2046 2,177,903 31,564,289 (28,424,843) 3,139,446 3,143,787 2,173,562 18,671 2047 2,194,558 30,841,004 (27,773,818) 3,067,186 3,071,784 2,189,959 16,398 2048 2,208,980 29,806,151 (26,842,200) 2,963,951 2,968,747 2,204,185 14,225 2049 2,221,293 28,470,031 (25,639,258) 2,830,773 2,835,702 2,216,365 12,180 2050 2,231,639 26,848,924 (24,179,646) 2,669,277 2,674,269 2,226,648 10,284 2051 2,240,168 24,972,749 (22,490,297) 2,482,451 2,487,427 2,235,193 8,545 2052 2,246,432 22,930,156 (20,650,250) 2,279,906 2,284,848 2,241,491 6,298 All figures are estimates. Actual results will depend upon mortality, interest rates and dividends. - 74 - Appendix B Scenario 1 - Constant Loan Interest Rate - March Issue Profit and Loss Statement (dollars in thousands except earnings per share) Pre-Tax Effect Tax Effect (A) (B) (C) (C1) (D) (E) (F) (G) (H) (I) (J) (K) Annual Accrued Deductible Pre-Tax Policy Admin. Net CSV Loan Loan Earnings Loan Fee Tax After-Tax After-Tax Annual Increase/ Interest Interest Death Admin. Effect Tax Tax Effect Earnings Earnings Year (Premium)* (Decrease) (Payment) (Payment) Benefits Fee A+B+C+D+E Credit Credit G+H Effect Per Share**[1] 1993 (108,573) 112,615 (11,191) (11,191) 3,250 (290) (4,188) 4,368 113 4,480 292 1994 (108,411) 119,289 (23,024) (23,024) 4,551 (289) (7,885) 8,988 113 9,101 1,217 1995 (106,499) 126,697 (35,570) (35,570) 4,791 (289) (10,869) 13,887 113 14,000 3,130 0.04 1996 20,817 (77) (35,489) (35,489) 5,064 (288) (9,972) 13,856 112 13,968 3,996 0.05 1997 20,730 (82) (35,401) (35,401) 5,371 (287) (9,669) 13,823 112 13,935 4,266 0.06 1998 20,601 (87) (35,308) (35,308) 5,715 (286) (9,366) 13,788 112 13,900 4,533 0.06 1999 20,426 (93) (35,208) (35,208) 6,097 (286) (9,063) 13,750 111 13,862 4,799 0.06 2000 (106,826) 140,401 (49,052) (49,052) 6,519 (285) (9,243) 19,159 111 19,270 10,027 0.13 2001 (106,493) 154,280 (64,224) (64,224) 6,985 (284) (9,736) 25,088 1ll 25,198 15,462 0.20 2002 (106,134) 169,781 (80,873) (80,973) 7,496 (283) (10,013) 31,595 110 31,705 21,692 0.28 2003 (105,748) 186,761 (99,131) (99,131) 8,055 (282) (10,345) 38,733 110 38,843 28,498 0.37 2004 (105,333) 205,194 (119,124) (119,124) 8,665 (281) (10,879) 46,551 110 46,661 35,782 0.47 2005 (104,886) 225,206 (140,990) (140,990) 9,331 (280) (11,619) 55,104 109 55,213 43,595 0.57 2006 (104,406) 247,089 (164,889) (164,889) 10,056 (278) (12,428) 64,456 109 64,564 52,136 0.68 2007 (76,351) 240,547 (187,967) (188,297) 10,839 (277) (13,208) 73,436 108 73,544 60,336 0.79 2008 167,368 (1,027) (186,925) (187,279) 11,678 (276) (9,182) 73,039 107 73,146 63,965 0.84 2009 165,195 (1,104) (185,806) (186,185) 12,58O (274) (9,409) 72,612 107 72,719 63,310 0.83 2010 159,327 (797) (184,604) (185,009) 16,696 (272) (9,651) 72,153 106 72,260 62,608 0.82 2011 156,940 (1,247) (183,315) (183,747) 18,005 (270) (9,888) 71,661 105 71,767 61,878 2012 154,027 (1,338) (181,935) (182,394) 19,409 (268) (10,104) 71,134 105 71,238 61,134 2013 144,448 5,292 (180,457) (180,946) 20,927 (266) (10,056) 70,569 104 70,673 60,616 0.79 2014 0 163,324 (195,835) (180,762) 21,690 (264) (11,085) 70,497 103 70,600 59,515 0.78 2015 0 177,742 (211,722) (179,242) 22,367 (262) (11,875) 69,905 102 70,007 58,132 2016 0 193,511 (228,882) (177,388) 22,927 (259) (12,704) 69,181 101 69,283 56,579 2017 0 210,606 (247,408) (175,397) 23,475 (256) (13,584) 68,405 100 68,505 54,921 2018 0 229,051 (267,377) (173,285) 24,195 (254) (14,384) 67,581 99 67,680 53,296 2019 0 248,689 (288,854) (171,050) 25,303 (250) (15,112) 66,710 98 66,807 51,696 0.67 2020 0 269,099 (311,842) (168,689) 27,085 (247) (15,905) 65,789 96 65,885 49,980 0.65 2021 0 290,908 (336,367) (166,200) 29,257 (244) (16,447) 64,818 95 64,913 48,467 0.63 2022 0 314,911 (362,586) (163,581) 31,579 (240) (16,336) 63,797 94 63,890 47,554 0.62 *Total annual premium less annual withdrawal. **Based on 76.6 million shares outstanding. Assumes 39 percent tax bracket. __________________________________ 1 Blank space indicates that there was no legible figure in underlying exhibit. Scenario 1 - Constant Loan Interest Rate - March Issue - 75 - Profit and Loss Statement (dollars in thousands except earnings per share) Pre-Tax Effect Tax Effect (A) (B) (C) (C1) (D) (E) (F) (G) (H) (I) (J) (K) Annual Accrued Deductible Pre-Tax Policy Admin. Net CSV Loan Loan Earnings Loan Fee Tax After-Tax After-Tax Annual Increase/ Interest Interest Death Admin. Effect Tax Tax Effect Earnings Earnings Year (Premium)* (Decrease) (Payment) (Payment) Benefits Fee A+B+C+D+E Credit Credit G+H Effect Per Share**[1] 2023 0 340,631 (390,574) (160,830) 34,055 (236) (16,124) 62,724 92 62,816 46,692 2024 0 368,031 (420,387) (157,945) 36,690 (232) (15,898) 61,599 91 61,689 45,791 2025 0 397,113 (452,064) (154,926) 39,518 (228) (15,662) 60,421 89 60,510 44,848 0.59 2026 0 427,884 (485,639) (151,772) 42,562 (224) (15,417) 59,191 87 59,278 43,862 0.57 2027 0 460,325 (521,120) (148,483) 45,852 (219) (15,163) 57,908 85 57,994 42,831 0.56 2028 0 494,399 (558,503) (145,060) 49,417 (214) (14,901) 56,573 84 56,657 41,756 0.55 2029 0 530,041 (597,761) (141,505) 53,298 (209) (14,632) 55,187 82 55,269 40,637 0.53 2030 0 567,297 (638,854) (137,821) 57,409 (204) (14,352) 53,750 80 53,830 39,478 0.52 2031 0 606,129 (681,731) (134,011) 61,738 (199) (14,063) 52,264 78 52,342 38,279 0.50 2032 0 646,466 (726,305) (130,083) 66,267 (193) (13,765) 50,732 75 50,808 37,043 0.48 2033 0 688,252 (772,492) (126,038) 70,970 (188) (13,457) 49,155 73 49,228 35,771 0.47 2034 0 731,334 (820,110) (121,884) 75,817 (182) (13,141) 47,535 71 47,606 34,465 0.45 2035 0 775,547 (868,971) (117,624) 80,784 (176) (12,816) 45,873 69 45,942 33,126 0.43 2036 0 820,751 (918,904) (113,262) 85,840 (170) (12,483) 44,172 66 44,238 31,755 0.41 2037 0 866,663 (969,570) (108,809) 90,931 (163) (12,139) 42,435 64 42,499 30,360 0.40 2038 0 913,101 (1,020,708) (104,268) 95,980 (157) (11,784) 40,665 61 40,726 28,942 0.38 2039 0 959,789 (1,071,984) (99,653) 100,929 (150) (11,417) 38,865 59 38,923 27,507 0.36 2040 0 1,006,430 (1,123,011) (94,973) 105,689 (144) (11,035) 37,040 56 37,096 26,060 0.34 2041 0 1,052,640 (1,173,299) (90,240) 110,158 (137) (10,638) 35,194 53 35,247 24,609 2042 0 1,097,999 (1,222,329) (85,467) 114,237 (130) (10,223) 33,332 51 33,383 23,160 2043 0 1,142,143 (1,269,661) (80,666) 117,847 (123) (9,794) 31,460 48 31,508 21,714 0.28 2044 0 1,184,472 (1,314,554) (75,854) 120,853 (116) (9,344) 29,583 45 29,628 20,284 0.26 2045 0 1,223,920 (1,356,161) (71,045) 123,447 (109) (8,902) 27,707 43 27,750 18,847 2046 0 1,259,769 (1,393,716) (66,257) 125,594 (102) (8,455) 25,840 40 25,880 17,425 2047 0 1,291,564 (1,426,758) (61,513) 127,285 (95) (8,006) 23,990 37 24,027 16,022 2048 0 1,318,398 (1,454,407) (56,838) 128,534 (88) (7,563) 22,167 34 22,201 14,638 2049 0 1,339,293 (1,475,845) (52,250) 129,481 (82) (7,153) 20,378 32 20,409 13,256 0.17 2050 0 1,353,960 (1,490,508) (47,770) 129,883 (75) (6,739) 18,630 29 18,660 11,921 0.16 2051 0 1,361,755 (1,498,003) (43,417) 129,910 (69) (6,407) 16,932 27 16,959 10,552 0.14 2052 0 1,361,650 (1,497,588) (39,210) 129,756 (62) (6,244) 15,292 24 15,316 9,072 0.12 *Total annual premium less annual withdrawal. **Based on 76.6 million shares outstanding. Assumes 39 percent tax bracket. __________________________________ 1 Blank space indicates that there was no legible figure in underlying exhibit. - 76 - Scenario 1 - Constant Loan Interest Rate - March Issue Cash Flow Detail (dollars in thousands) Corporate Cash Outflow Corporate Cash Inflow Net Cash Flow Surplus (A) (B) (C) (D) (E) (F) (G) (H) (I) (J) (K) Cumulative Cash After-Tax Policy Loan Tax-Free Net Cumulative Flow Cumulative Loan Admin. Tax Policy Policy Death Cash Cash at 4.35% Net Year Premium Interest Fee Savings Loan Withdrawal Benefits Clow Flow Pre-Tax Equity[1] 1993 (108,573) 0 (177) 4,368 101,328 0 3,250 196 196 98 96 1994 (108,411) (11,191) (176) 8,988 107,413 0 4,551 1,174 1,370 1,122 139 1995 (108,188) (23,024) (176) 13,887 114,126 1,689 4,791 3,105 4,475 4,087 164 1996 (107,951) (35,570) (176) 13,856 0 128,768 5,064 3,991 8,466 8,040 169 1997 (107,698) (35,489) (175) 13,823 0 128,427 5,371 4,260 12,727 12,371 175 1998 (107,427) (35,401) (175) 13,788 0 128,028 5,715 4,527 17,254 17,089 181 1999 (107,137) (35,308) (174) 13,750 0 127,564 6,097 4,792 22,045 22,202 188 2000 (106,826) (35,208) (174) 19,159 126,528 0 6,519 9,999 32,044 32,722 216 2001 (106,493) (49,052) (173) 25,088 139,077 0 6,985 15,431 47,475 49,017 247 2002 (106,134) (64,224) (173) 31,595 153,098 0 7,496 21,658 69,134 72,052 281 2003 (105,748) (80,873) (172) 38,733 168,466 0 8,055 28,460 97,594 102,591 318 2004 (105,333) (99,131) (171) 46,551 185,160 0 8,665 35,742 133,336 141,317 359 2005 (104,886) (119,124) (171) 55,104 203,297 0 9,331 43,551 176,887 188,984 402 2006 (104,406) (140,990) (170) 64,456 223,142 0 10,056 52,088 228,976 246,569 450 2007 (76,351) (164,889) (169) 73,436 217,421 0 10,839 60,288 289,263 313,996 498 2008 0 (187,967) (168) 73,039 0 167,368 11,678 63,950 353,213 387,000 512 2009 0 (186,925) (167) 72,612 0 165,195 12,580 63,294 416,508 461,294 527 2010 0 (185,806) (166) 72,153 0 159,327 16,696 62,204 478,712 536,425 932 2011 0 (184,604) (165) 71,661 0 156,940 18,005 61,837 540,548 613,196 973 2012 0 (183,315) (164) 71,134 0 154,027 19,409 61,091 601,639 691,263 1,017 2013 0 (181,935) (162) 70,569 0 144,448 20,927 53,847 655,486 763,984 7,786 2014 0 (180,457) (161) 70,497 154,593 0 21,690 66,161 721,647 851,306 1,140 2015 0 (195,835) (160) 69,905 161,737 0 22,367 58,013 779,660 932,617 2016 0 (211,722) (158) 69,181 176,233 0 22,927 56,461 836,121 1,014,532 2017 0 (228,882) (156) 68,405 192,048 0 23,475 54,889 891,011 1,097,047 2018 0 (247,408) (155) 67,581 209,115 0 24,195 53,328 944,339 1,180,188 1,377 2019 0 (267,377) (153) 66,710 227,389 0 25,303 51,872 996,211 1,264,075 1,200 2020 0 (288,854) (151) 65,789 246,381 0 27,085 50,250 1,046,461 1,348,549 930 2021 0 (311,842) (149) 64,818 266,366 0 29,257 48,450 1,094,911 1,433,440 947 2022 0 (336,367) (146) 63,797 288,659 0 31,579 47,520 1,142,432 1,519,651 980 Assumes 39 percent tax bracket. All figures are estimates. Actual results will depend upon mortality, interest rates and dividends. __________________________________ 1 Blank space indicates that there was no legible figure in underlying exhibit. - 77 - Scenario 1 - Constant Loan Interest Rate - March Issue Cash Flow Detail (dollars in thousands) Corporate Cash Outflow Corporate Cash Inflow Net Cash Flow Surplus (A) (B) (C) (D) (E) (F) (G) (H) (I) (J) (K) Cumulative Cash After-Tax Policy Loan Tax-Free Net Cumulative Flow Cumulative Loan Admin. Tax Policy Policy Death Cash Cash at 4.35% Net Year Premium Interest Fee Savings Loan Withdrawal Benefits Clow Flow Pre-Tax Equity[1] 2023 0 (362,586) (144) 62,724 312,610 0 34,055 46,659 1,189,090 1,607,284 1,013 2024 0 (390,574) (142) 61,599 338,182 0 36,690 45,755 1,234,846 1,696,336 1,049 2025 0 (420,387) (139) 60,421 365,397 0 39,518 44,809 1,279,655 1,786,798 1,088 2026 0 (452,064) (136) 59,191 394,268 0 42,562 43,820 1,323,475 1,878,662 1,129 2027 0 (485,639) (134) 57,908 424,799 0 45,852 42,787 1,366,262 1,971,918 1,174 2028 0 (521,120) (131) 56,573 456,970 0 49,417 41,709 1,407,972 2,066,555 1,221 2029 0 (558,503) (128) 55,187 490,734 0 53,298 40,588 1,448,560 2,162,562 1,269 2030 0 (597,761) (125) 53,750 526,152 0 57,409 39,426 1,487,986 2,259,931 1,322 2031 0 (638,854) (121) 52,264 563,196 0 61,738 38,223 1,526,209 2,358,655 1,378 2032 0 (681,731) (118) 50,732 601,833 0 66,267 36,983 1,563,192 2,458,730 1,437 2033 0 (726,305) (114) 49,155 642,002 0 70,970 35,708 1,598,900 2,560,155 1,500 2034 0 (772,492) (111) 47,535 683,650 0 75,817 34,398 1,633,298 2,662,930 1,567 2035 0 (820,110) (107) 45,873 726,616 0 80,784 33,056 1,666,354 2,767,056 1,636 2036 0 (868,971) (103) 44,172 770,746 0 85,840 31,683 1,698,036 2,872,538 1,709 2037 0 (918,904) (100) 42,435 815,922 0 90,931 30,284 1,728,320 2,979,385 1,785 2038 0 (969,570) (96) 40,665 861,884 0 95,980 28,863 1,757,184 3,087,615 1,864 2039 0 (1,020,708) (92) 38,865 908,432 0 100,929 27,426 1,784,610 3,197,247 1,944 2040 0 (1,071,984) (88) 37,040 955,322 0 105,689 25,978 1,810,588 3,308,313 2,026 2041 0 (1,123,011) (83) 35,194 1,002,270 0 110,158 24,527 1,835,115 3,420,851 2,109 2042 0 (1,173,299) (79) 33,332 1,048,887 0 114,237 23,078 1,858,192 3,534,909 2,191 2043 0 (1,222,329) (75) 31,460 1,094,731 0 117,847 21,634 1,879,827 3,650,541 2,270 2044 0 (1,269,661) (71) 29,583 1,139,503 0 120,853 20,208 1,900,034 3,767,815 2,346 2045 0 (1,314,554) (67) 27,707 1,182,248 0 123,447 18,783 1,918,817 3,886,781 2,411 2046 0 (1,356,161) (62) 25,840 1,222,163 0 125,594 17,375 1,936,192 4,007,510 2047 0 (1,393,716) (58) 23,990 1,258,488 0 127,285 15,988 1,952,180 4,130,077 2048 0 (1,426,758) (54) 22,167 1,290,735 0 128,534 14,623 1,966,803 4,254,559 2049 0 (1,454,407) (50) 20,378 1,317,860 0 129,481 13,262 1,980,064 4,381,015 2,505 2050 0 (1,475,845) (46) 18,630 1,339,314 0 129,883 11,937 1,992,002 4,509,544 2,489 2051 0 (1,490,508) (42) 16,932 1,354,286 0 129,910 10,579 2,002,580 4,640,170 2,462 2052 0 (1,498,003) (38) 15,292 1,362,102 0 129,756 9,109 2,011,689 4,772,838 2,425 Assumes 39 percent tax bracket. All figures are estimates. Actual results will depend upon mortality, interest rates and dividends. __________________________________ 1 Blank space indicates that there was no legible figure in underlying exhibit. - 78 - Scenario 1 - Constant Loan Interest Rate - March Issue Balance Sheet Summary (dollars in thousands) (A) (B) (C) (D) (E) (F) (G) (H) Gross Insurance Net Accrued Retained Annual Cash Cash Surrender Outstanding Cash Surrender Loan Asset Earnings Impact on [2] Year Amount Value (Loan) Value Interest Balance[1] Gain/(Loss) Earnings 1993 196 112,471 (101,184) 11,287 11,191 22,478 292 1994 1,370 231,340 (208,177) 23,163 23,024 46,187 1,509 1995 4,475 357,345 (321,611) 35,734 35,570 71,304 4,639 3,130 1996 8,466 356,530 (320,873) 35,658 35,489 71,146 8,635 3,996 1997 12,727 355,660 (320,084) 35,576 35,401 70,977 12,901 4,266 1998 17,254 354,728 (319,239) 35,489 35,308 70,797 17,435 4,533 1999 22,045 353,729 (318,333) 35,396 35,208 70,603 22,233 4,799 2000 32,044 492,780 (443,511) 49,268 49,052 98,321 32,260 10,027 2001 47,475 645,163 (580,691) 64,471 64,224 128,696 47,722 15,462 2002 69,134 812,376 (731,222) 81,154 80,873 162,027 69,415 21,692 2003 97,594 995,749 (896,300) 99,449 99,131 198,580 97,912 28,498 2004 133,336 1,196,556 (1,077,073) 119,483 119,124 238,607 133,695 35,782 2005 176,887 1,416,165 (1,274,773) 141,392 140,990 282,382 177,289 43,595 2006 228,976 1,656,195 (1,490,857) 165,338 164,889 330,227 229,425 52,136 2007 289,263 1,887,983 (1,699,519) 188,464 187,967 376,431 289,761 60,336 2008 353,213 1,877,537 (1,690,100) 187,437 186,925 374,362 353,725 63,965 2009 416,508 1,866,313 (1,679,980) 186,333 185,806 372,139 417,035 63,310 2010 478,712 1,854,651 (1,669,115) 185,536 184,604 370,140 479,643 62,608 2011 540,548 1,841,752 (1,657,463) 184,289 183,315 367,604 541,522 2012 601,639 1,827,931 (1,644,980) 182,951 181,935 364,886 602,656 2013 655,486 1,819,864 (1,631,621) 188,244 180,457 368,701 663,272 60,616 2014 721,647 1,967,636 (1,770,661) 196,975 195,835 392,811 722,787 59,515 2015 779,660 2,127,281 (1,914,300) 212,981 211,722 424,702 780,919 2016 836,121 2,299,721 (2,069,461) 230,259 228,882 459,142 837,498 2017 891,011 2,485,785 (2,236,968) 248,817 247,408 496,225 892,420 2018 944,339 2,686,271 (2,417,518) 268,754 267,377 536,131 945,715 53,296 2019 996,211 2,901,756 (2,611,702) 290,054 288,854 578,907 997,411 51,696 2020 1,046,461 3,132,325 (2,819,553) 312,772 311,842 624,614 1,047,391 49,980 2021 1,094,911 3,378,611 (3,041,297) 337,314 336,367 673,681 1,095,858 48,467 2022 1,142,432 3,641,933 (3,278,367) 363,566 362,586 726,152 1,143,412 47,554 All figures are estimates. Actual results will depend upon mortality, interest rates and dividends 1 Note: The yearly Asset Balances shown above in column F are the sum of the amounts in columns D and E. The proper asset balances should be the Net Cash Surrender Values shown in column D minus the Accrued Loan Interest shown in column E. For example, for 1993, $11,287 minus $11,191 equals $96. $96 plus the cash flow of $196 (column A) equals the total gain shown in Column G. 2 Blank space indicates that there was no legible figure in underlying exhibit. - 79 - Scenario 1 - Constant Loan Interest Rate - March Issue Balance Sheet Summary (dollars in thousands) (A) (B) (C) (D) (E) (F) (G) (H) Gross Insurance Net Accrued Retained Annual Cash Cash Surrender Outstanding Cash Surrender Loan Asset Earnings Impact on [2] Year Amount Value (Loan) Value Interest Balance[1] Gain/(Loss) Earnings 2023 1,189,090 3,923,010 (3,531,423) 391,587 390,574 782,161 1,190,104 2024 1,234,846 4,222,408 (3,800,972) 421,436 420,387 841,823 1,235,895 2025 1,279,655 4,540,559 (4,087,407) 453,152 452,064 905,216 1,280,743 44,848 2026 1,323,475 4,877,748 (4,390,980) 486,768 485,639 972,406 1,324,605 43,862 2027 1,366,262 5,234,121 (4,711,827) 522,294 521,120 1,043,414 1,367,436 42,831 2028 1,407,972 5,609,615 (5,049,892) 559,723 558,503 1,118,226 1,409,192 41,756 2029 1,448,560 6,003,965 (5,404,934) 599,030 597,761 1,196,791 1,449,829 40,637 2030 1,487,986 6,416,810 (5,776,634) 640,176 638,854 1,279,029 1,489,307 39,478 2031 1,526,209 6,847,570 (6,164,461) 683,109 681,731 1,364,840 1,527,586 38,279 2032 1,563,192 7,295,469 (6,567,727) 727,742 726,305 1,454,047 1,564,629 37,043 2033 1,598,900 7,759,356 (6,985,364) 773,992 772,492 1,564,484 1,600,400 35,771 2034 1,633,298 8,237,806 (7,416,130) 821,676 820,110 1,641,786 1,634,864 34,465 2035 1,666,354 8,728,882 (7,858,275) 870,608 868,971 1,739,579 1,667,990 33,126 2036 1,698,036 9,230,346 (8,309,732) 920,613 918,904 1,839,517 1,699,745 31,755 2037 1,728,320 9,739,681 (8,768,326) 971,355 969,570 1,940,925 1,730,105 30,360 2038 1,757,184 10,253,679 (9,231,107) 1,022,571 1,020,708 2,043,279 1,759,047 28,942 2039 1,784,610 10,769,072 (9,695,144) 1,073,928 1,071,984 2,145,913 1,786,554 27,507 2040 1,810,588 11,281,885 (10,156,848) 1,125,037 1,123,011 2,248,048 1,812,614 26,060 2041 1,835,115 11,787,573 (10,612,166) 1,175,408 1,173,299 2,348,707 1,837,223 2042 1,858,192 12,281,040 (11,056,521) 1,224,520 1,222,329 2,446,849 1,860,383 2043 1,879,827 12,756,559 (11,484,628) 1,271,931 1,269,661 2,541,592 1,882,097 21,714 2044 1,900,034 13,207,807 (11,890,907) 1,316,900 1,314,554 2,631,454 1,902,381 20,284 2045 1,918,817 13,626,925 (12,268,353) 1,358,572 1,356,161 2,714,733 1,921,228 2046 1,936,192 14,006,378 (12,610,200) 1,396,178 1,393,716 2,789,894 1,938,654 2047 1,952,180 14,338,988 (12,909,734) 1,429,254 1,426,758 2,856,012 1,954,675 2048 1,966,803 14,617,963 (13,161,046) 1,456,917 1,454,407 2,911,324 1,969,313 14,638 2049 1,980,064 14,835,464 (13,357,114) 1,478,350 1,475,845 2,954,194 1,982,570 13,256 2050 1,992,002 14,985,118 (13,492,121) 1,492,997 1,490,508 2,983,505 1,994,490 11,921 2051 2,002,580 15,061,908 (13,561,443) 1,500,465 1,498,003 2,998,469 2,005,043 10,552 2052 2,011,689 15,061,904 (13,561,891) 1,500,014 1,497,588 2,997,602 2,014,114 9,072 All figures are estimates. Actual results will depend upon mortality, interest rates and dividends. __________________________________ 1 Note: The yearly Asset Balances shown above in column F are the sum of the amounts in columns D and E. The proper asset balances should be the Net Cash Surrender Values shown in column D minus the Accrued Loan Interest shown in column E. For example, for 2023, $391,587 minus $390,574 equals $1,013. $1,013 plus the cash flow of $1,189,090 (column A) equals the total gain shown in Column G. 2 Blank space indicates that there was no legible figure in underlying exhibit. Scenario 1 - Constant Loan Interest Rate - March Issue - 80 - Effect on Financial Ratios (dollars in thousands) (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) Adjusted Operating Current Pre-Tax Adjusted Expense as COLI Projected Operating COLI Operating a % of Pre-Tax Taxes Tax Adjusted Adjusted Year Sales Expense Expense Expense Sales Earnings Paid at 39% Savings Taxes Paid Tax Rate[1] 1993 10,800,000 2,160,000 4,188 2,164,188 20.04% 337,500 131,625 4,480 127,145 37.67% 1994 11,232,000 2,246,400 7,885 2,254,285 20.07% 351,000 136,890 9,101 127,789 36.41% 1995 11,681,280 2,336,256 10,869 2,347,125 20.09% 365,040 142,366 14,000 128,366 35.16% 1996 12,148,531 2,429,706 9,972 2,439,678 20.08% 379,642 148,060 13,968 134,092 35.32% 1997 12,634,472 2,526,894 9,669 2,536,564 20.08% 394,827 153,983 13,935 140,047 35.47% 1998 13,139,851 2,627,970 9,366 2,637,337 20.07% 410,620 160,142 13,900 146,242 35.61% 1999 13,665,445 2,733,089 9,063 2,742,152 20.07% 427,045 166,548 13,862 152,686 35.75% 2000 14,212,063 2,842,413 9,243 2,851,656 20.07% 444,127 173,210 19,270 153,939 34.66% 2001 14,780,546 2,956,109 9,736 2,965,845 20.07% 461,892 180,138 25,198 154,939 33.54% 2002 15,371,768 3,074,354 10,013 3,084,366 20.07% 480,368 187,343 31,705 155,638 32.40% 2003 15,986,638 3,197,328 10,345 3,207,673 20.06% 499,582 194,837 38,843 155,995 31.22% 2004 16,626,104 3,325,221 10,879 3,336,099 20.07% 519,566 202,631 46,661 155,970 30.02% 2005 17,291,148 3,458,230 11,619 3,469,848 20.07% 540,348 210,736 55,213 155,522 28.78% 2006 17,982,794 3,596,559 12,428 3,608,987 20.07% 561,962 219,165 64,564 154,601 27.51% 2007 18,702,106 3,740,421 13,208 3,753,629 20.07% 584,441 227,932 73,544 154,388 26.42% 2008 19,450,190 3,890,038 9,182 3,899,220 20.05% 607,818 237,049 73,146 163,903 26.97% 2009 20,228,197 4,045,639 9,409 4,055,049 20.05% 632,131 246,531 72,719 173,812 27.50% 2010 21,037,325 4,207,465 9,651 4,217,116 20.05% 657,416 256,392 72,260 184,133 28.01% 2011 21,878,818 4,375,764 9,888 4,385,652 20.05% 683,713 266,648 71,767 194,881 28.50% 2012 22,753,971 4,550,794 10,104 4,560,899 20.04% 711,062 277,314 71,238 206,076 28.98% 2013 23,664,130 4,732,826 10,056 4,742,882 20.04% 739,504 288,407 70,673 217,734 29.44% 2014 24,610,695 4,922,139 11,085 4,933,224 20.05% 769,084 299,943 70,600 229,343 29.82% 2015 25,595,123 5,119,025 11,875 5,130,899 20.05% 799,848 311,941 70,007 241,934 2016 26,618,928 5,323,786 12,704 5,336,489 20.05% 831,841 324,418 69,283 255,136 2017 27,683,685 5,536,737 13,584 5,550,321 20.05% 865,115 337,395 68,505 268,890 2018 28,791,032 5,758,206 14,384 5,772,591 20.05% 899,720 350,891 67,680 283,211 31.48% 2019 29,942,674 5,988,535 15,112 6,003,647 20.05% 935,709 364,926 66,807 298,119 31.86% 2020 31,140,381 6,228,076 15,905 6,243,981 20.05% 973,137 379,523 65,885 313,638 32.23% 2021 32,385,996 6,477,199 16,447 6,493,646 20.05% 1,012,062 394,704 64,913 329,791 32.59% 2022 33,681,436 6,736,287 16,336 6,752,623 20.05% 1,052,545 410,492 63,890 346,602 32.93% Assumes 39 percent tax bracket. All figures are estimates. Actual results will depend upon mortality, interest rates and dividends. __________________________________ 1 Blank space indicates that there was no legible figure in underlying exhibit. Scenario 1 - Constant Loan Interest Rate - March Issue - 81 - Effect on Financial Ratios (dollars in thousands) (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) Adjusted Operating Current Pre-Tax Adjusted Expense as COLI Projected Operating COLI Operating a % of Pre-Tax Taxes Tax Adjusted Adjusted Year Sales Expense Expense Expense Sales Earnings Paid at 39% Savings Taxes Paid Tax Rate[1] 2023 35,028,693 7,005,739 16,124 7,021,863 20.05% 1,094,647 426,912 62,816 364,096 33.26% 2024 36,429,841 7,285,968 15,898 7,301,866 20.04% 1,138,433 443,989 61,689 382,300 33.58% 2025 37,887,034 7,577,407 15,662 7,593,069 20.04% 1,183,970 461,748 60,510 401,238 33.89% 2026 39,402,516 7,880,503 15,417 7,895,920 20.04% 1,231,329 480,218 59,278 420,940 34.19% 2027 40,978,616 8,195,723 15,163 8,210,886 20.04% 1,280,582 499,427 57,994 441,433 34.47% 2028 42,617,761 8,523,552 14,901 8,538,453 20.03% 1,331,805 519,404 56,657 462,747 34.75% 2029 44,322,472 8,864,494 14,632 8,879,126 20.03% 1,385,077 540,180 55,269 484,912 35.01% 2030 46,095,370 9,219,074 14,352 9,233,426 20.03% 1,440,480 561,787 53,830 507,958 35.26% 2031 47,939,185 9,587,837 14,063 9,601,900 20.03% 1,498,100 584,259 52,342 531,917 35.51% 2032 49,856,753 9,971,351 13,765 9,985,115 20.03% 1,558,024 607,629 50,808 556,822 35.74% 2033 51,851,023 10,370,205 13,457 10,383,662 20.03% 1,620,344 631,934 49,228 582,706 35.96% 2034 53,925,064 10,785,013 13,141 10,798,154 20.02% 1,685,158 657,212 47,606 609,606 36.18% 2035 56,082,066 11,216,413 12,816 11,229,229 20.02% 1,752,565 683,500 45,942 637,558 36.38% 2036 58,325,349 11,665,070 12,483 11,677,553 20.02% 1,822,667 710,840 44,238 666,602 36.57% 2037 60,658,363 12,131,673 12,139 12,143,812 20.02% 1,895,574 739,274 42,499 696,775 36.76% 2038 63,084,697 12,616,939 11,784 12,628,723 20.02% 1,971,397 768,845 40,726 728,119 36.93% 2039 65,608,085 13,121,617 11,417 13,133,034 20.02% 2,050,253 799,599 38,923 760,675 37.10% 2040 68,232,409 13,646,482 11,035 13,657,517 20.02% 2,132,263 831,582 37,096 794,487 37.26% 2041 70,961,705 14,192,341 10,638 14,202,979 20.01% 2,217,553 864,846 35,247 829,599 37.41% 2042 73,800,173 14,760,035 10,223 14,770,258 20.01% 2,306,255 899,440 33,383 866,057 37.55% 2043 76,752,180 15,350,436 9,794 15,360,230 20.01% 2,398,506 935,417 31,508 903,910 37.69% 2044 79,822,267 15,964,453 9,344 15,973,798 20.01% 2,494,446 972,834 29,628 943,206 37.81% 2045 83,015,158 16,603,032 8,902 16,611,934 20.01% 2,594,224 1,011,747 27,750 983,997 2046 86,335,764 17,267,153 8,455 17,275,608 20.01% 2,697,993 1,052,217 25,880 1,026,337 2047 89,789,195 17,957,839 8,006 17,965,845 20.01% 2,805,912 1,094,306 24,027 1,070,279 38.14% 2048 93,380,763 18,676,153 7,563 18,683,716 20.01% 2,918,149 1,138,078 22,201 1,115,877 38.24% 2049 97,115,993 19,423,199 7,153 19,430,352 20.01% 3,034,875 1,183,601 20,409 1,163,192 38.33% 2050 101,000,633 20,200,127 6,739 20,206,866 20.01% 3,156,270 1,230,945 18,660 1,212,286 38.41% 2051 105,040,658 21,008,132 6,407 21,014,539 20.01% 3,282,521 1,280,183 16,959 1,263,224 38.48% 2052 109,242,285 21,848,457 6,244 21,854,701 20.01% 3,413,821 1,331,390 15,316 1,316,074 38.55% Assumes 39 percent tax bracket. All figures are estimates. Actual results will depend upon mortality, interest rates and dividends. __________________________________ 1 Blank space indicates that there was no legible figure in underlying exhibit. Footnotes [1] - Winn-Dixie makes deposits and pays loan interest to insurance carrier. [2] - Winn-Dixie receives withdrawals, loans and death proceeds from the insurance carrier. [3] - Winn-Dixie receives a tax deduction for loan interest paid. The memorandum next explained the difference between the proposed broad-based COLI pool and petitioner's then existing leveraged COLI program being used to fund the MSP. The memorandum stated: Winn-Dixie is familiar with leveraged COLI and particularly with the tax arbitrage created when deductible policy loan interest is paid to finance non- taxable policy gains. Winn-Dixie's existing leveraged COLI policies provide this arbitrage and, having been purchased before passage of the 1986 Tax Reform Act, provide it beyond the $50,000 cap applicable to newer policies.