1998 Tax Ct. Memo LEXIS 463 | Tax Ct. | 1998
Decision will be entered under Rule 155.
MEMORANDUM FINDINGS OF FACT AND OPINION
COLVIN, JUDGE: Respondent determined that petitioner had deficiencies in income tax of $ 5,280,264 for 1991 and $ 1,905,200 for 1992.
The sole issue for decision is whether petitioner may deduct $ 45,650,249 for its reserves for discounted unpaid losses and loss adjustment expenses for 1991 and $ 49,418,509 for 1992. We hold that it may.
Section references are to the Internal Revenue Code. Unless otherwise indicated, Rule references are to the Tax Court Rules of Practice and Procedure.
Utah Medical Insurance Association (referred to as petitioner) is a property and casualty insurance company the principal place of business of which is in Salt Lake City, Utah. Petitioner provides medical malpractice liability insurance for physicians in Utah, Montana, and Idaho. Medical malpractice liability insurance indemnifies a physician against medical professional liability 1998 Tax Ct. Memo LEXIS 463" label="1998 Tax Ct. Memo LEXIS 463" no-link"="" number="2" pagescheme="<span class=">1998 Tax Ct. Memo LEXIS 463">*464 claims for damages brought as a result of the provision of, or the failure to provide, medical services. Petitioner is, and during the years in issue was, taxed as a property and casualty insurance company under sections 831-835. Petitioner is managed by its board of directors, which is composed of 12 of petitioner's policyholder-physicians.
In the 1960's, medical malpractice liability insurance was generally provided by commercial insurance companies. Beginning around 1965, commercial insurance companies experienced large underwriting losses as a result of a rapid increase in medical malpractice claims and litigation. As a result, they raised rates, e.g., 400-600 percent in California from 1965 to 1971, to cover their losses. When rate increases failed to keep pace with continued ad- verse loss experience, many commercial insurers stopped issuing medical malpractice insurance. As a result, State medical societies formed physician-owned medical malpractice insurance companies to offer medical malpractice insurance to their members.
In the early 1970's, the Utah Medical Association (UMA), the leading professional association for doctors in Utah, endorsed 1998 Tax Ct. Memo LEXIS 463" label="1998 Tax Ct. Memo LEXIS 463" no-link"="" number="3" pagescheme="<span class=">1998 Tax Ct. Memo LEXIS 463">*465 Aetna Life and Casualty Insurance Co. (Aetna) as the preferred malpractice carrier in Utah. Aetna, which wrote most of the medical malpractice insurance in Utah during the 1970's, increased rates several times in the late 1970's.
In response to Aetna's rate increases, about 900 doctors who were members of UMA formed petitioner as an unincorporated inter- insurance exchange 1 or reciprocal company in November 1978. They executed subordinated loans which gave petitioner an initial capitalization of $ 2.2 million. Shortly thereafter, Aetna withdrew from the insurance market in Utah. Petitioner became the principal medical malpractice insurer in Utah.
From 1978 to 1981, petitioner wrote medical malpractice insurance only in Utah. In 1982, petitioner began issuing insurance policies to physicians practicing in Montana, and in 1991 to physicians practicing in Idaho. During the years in issue, about 85- 90 percent of petitioner's insurance policies were issued to doctors practicing in Utah.
During those years, petitioner primarily wrote medical malpractice 1998 Tax Ct. Memo LEXIS 463" label="1998 Tax Ct. Memo LEXIS 463" no-link"="" number="4" pagescheme="<span class=">1998 Tax Ct. Memo LEXIS 463">*466 liability insurance and also wrote a small amount of general liability insurance for its covered physicians.
From 1984 to 1992, the number of doctors insured by petitioner increased as follows:
| Number of | |
| Year | Insureds |
| 1984 | 1,369 |
| 1985 | 1,432 |
| 1986 | 1,482 |
| 1987 | 1,630 |
| 1988 | 1,699 |
| 1989 | 1,819 |
| 1990 | 1,814 |
| 1991 | 1,920 |
| 1992 | 2,002 |
Petitioner is principally regulated by the Utah Department of Insurance (UDI). Petitioner maintained its books and records in accordance with UDI requirements and filed annual statements with UDI. Petitioner prepared each annual statement in the format prescribed by the National Association of Insurance Commissioners (NAIC), a voluntary association of State insurance commissioners.
Insurance companies use "statutory accounting" principles to prepare their annual statements. Statutory accounting principles are conservative and focus on maintaining the solvency of an insurance company to protect insurance consumers.
UDI required that annual statements due after December 31, 1991, be accompanied by an actuarial opinion concerning the reasonableness of the insurance company's reserves. The actuarial firm of Tillinghast Towers Perrin (Tillinghast) 1998 Tax Ct. Memo LEXIS 463" label="1998 Tax Ct. Memo LEXIS 463" no-link"="" number="5" pagescheme="<span class=">1998 Tax Ct. Memo LEXIS 463">*467 certified to UDI that petitioner's reserves for unpaid losses shown on its 1991 and 1992 annual statements were computed in accordance with accepted loss reserving standards and were fairly stated in accordance with sound loss reserving principles, were based on factors relevant to policy provisions, met the requirements of the insurance laws of the State of Utah, and provided sufficiently for all of petitioner's unpaid loss and loss expense obligations.
On their annual statements, property and casualty insurers are required to report estimates of amounts they expect to pay for losses 2 that have already occurred (unpaid losses) and related loss adjustment expenses. These estimates are known as reserves for unpaid losses. A property and casualty company's loss reserve is the amount that is needed to make all future payments on claims that have already been incurred.
The reserve for unpaid losses 3 includes all incurred losses. Incurred losses include insured events for which a claim has been filed (reported losses), and insured events for which no claim has been filed with the company; i.e., incurred but not reported (IBNR) losses.
The first step in estimating loss reserves is to estimate the total amount that will ultimately be paid for a coverage year for all claims existing on a given date (ultimate losses). Ultimate losses equal paid losses plus an estimate of unpaid losses at the end of the year. Unpaid losses are generally estimated not later than the year in which the insured event giving rise to the loss occurred (the accident year). 4 In estimating its ultimate losses each year, 1998 Tax Ct. Memo LEXIS 463" label="1998 Tax Ct. Memo LEXIS 463" no-link"="" number="7" pagescheme="<span class=">1998 Tax Ct. Memo LEXIS 463">*469 petitioner adjusted its prior estimates of earlier years' ultimate losses to factor in its loss experience.
Petitioner's president, Martin J. Oslowski (Oslowski), recommended an amount to report as annual statement unpaid losses to the board of directors. Oslowski was petitioner's claims manager before he became president in December 1986.
Petitioner wrote only one line of insurance, and thus petitioner could not offset reserve deficiencies with surpluses in another line as multiline companies could do.
Medical malpractice insurance may be written on an "occurrence" basis or on a "claims-made" basis. An occurrence basis policy covers losses that occur within the policy period whenever reported. A claims-made basis policy covers only losses from occurrences during the policy period (or a previous policy period) for which claims are made or which are otherwise reported during the period. From 1978 to 1985, petitioner and most of the medical malpractice insurance industry offered only 1998 Tax Ct. Memo LEXIS 463" label="1998 Tax Ct. Memo LEXIS 463" no-link"="" number="8" pagescheme="<span class=">1998 Tax Ct. Memo LEXIS 463">*470 occurrence basis policies.
In 1984, petitioner's financial position was tenuous because its losses in the early 1980's were significantly larger than its reserves. Petitioner believed that a switch from occurrence to claims-made basis policies would improve its financial condition. In 1985, based on the recommendation of Tillinghast, petitioner began offering claims-made instead of occurrence basis policies to its physicians. Petitioner's claims-made policies had a 1-year term and an anniversary date 5 of January 1. Occurrence basis medical malpractice insurance policies are "long-tailed" because it can take 10 years or more for claims to be received and resolved.
Petitioner has hired outside actuaries to perform all of its actuarial services since it was formed.
1. MILLIMAN & ROBERTSON
From 1978 to 1985, the actuarial firm of Milliman Robertson (M&R) provided actuarial services to petitioner to help it estimate its annual statement unpaid losses.
Petitioner followed M&R's recommendations. However, M&R underestimated petitioner's 1998 Tax Ct. Memo LEXIS 463" label="1998 Tax Ct. Memo LEXIS 463" no-link"="" number="9" pagescheme="<span class=">1998 Tax Ct. Memo LEXIS 463">*471 unpaid losses. As this unfavorable trend emerged, petitioner gradually increased its estimates of unpaid losses each year.
To improve its financial condition, in 1984 petitioner asked UDI for permission to discount its loss reserves for the occurrence basis years. UDI approved petitioner's request. However, UDI required petitioner to begin filing quarterly statements and to provide UDI with loss and investment information. UDI, in effect, began to oversee petitioner's operations.
2. TILLINGHAST
In 1984, petitioner hired Tillinghast to determine whether petitioner had sufficient assets and surplus to meet its liabilities. Tillinghast concluded that petitioner did not.
3. TILLINGHAST'S LOSS RESERVE REVIEWS AND RATE REVIEWS
Tillinghast began preparing loss reserve reviews 6 and rate reviews 7 for petitioner at the end of 1985.
Sometimes actuaries estimate ultimate losses as a range with high and low bounds (a "bounded range") instead of as a single number (a "point estimate"). 1998 Tax Ct. Memo LEXIS 463" label="1998 Tax Ct. Memo LEXIS 463" no-link"="" number="10" pagescheme="<span class=">1998 Tax Ct. Memo LEXIS 463">*472 Actuarial Standard of Practice No. 9 states: "The uncertainty inherent in the estimation of required provisions for unpaid losses or loss adjustment expenses implies that a range of reserves can be actuarially sound." Tillinghast estimated petitioner's ultimate losses within a bounded range.
Beginning in 1989, James Hurley (Hurley), an actuary employed by Tillinghast, prepared annual rate reviews and semiannual loss reserve reviews for petitioner. In preparing his reserve reviews, Hurley received information from petitioner about its paid losses and case reserves. 8 Petitioner's claims investigators generally established petitioner's case reserves based on their initial impression of each claim and revised them as they acquired more information.
The high end of Tillinghast's estimate of petitioner's ultimate losses in 1986 and the high end reestimates of its ultimate losses in later years were as follows:
| Tillinghast's High End Estimates Of Petitioner's | ||||||
| 1Ultimate Net Losses & Allocated Loss Adjustment Expenses | ||||||
| (in thousands) | ||||||
| Year | 1986 | 1987 | 1988 | 1989 | 1990 | 1991 |
| 1986 | $ 3,544 | $ 2,481 | $ 2,293 | $ 2,186 | $ 2,700 | $ 2,700 |
| 1987 | -- | 8,450 | 6,840 | 6,655 | 6,300 | 7,250 |
| 1988 | -- | -- | 12,791 | 11,372 | 9,900 | 9,250 |
| 1989 | -- | -- | -- | 13,782 | 12,500 | 12,250 |
| 1990 | -- | -- | -- | -- | 16,000 | 15,250 |
| 1991 | -- | -- | -- | -- | -- | 15,500 |
| 1992 | -- | -- | -- | -- | -- | -- |
| 1993 | -- | -- | -- | -- | -- | -- |
| 1994 | -- | -- | -- | -- | -- | -- |
| IBNR | 665 | 1,476 | 1,777 | 1,361 | 1.625 | |
| Total | 3,544 | 11,596 | 23,400 | 35,772 | 48,761 | 63,825 |
Table 1998 Tax Ct. Memo LEXIS 463" label="1998 Tax Ct. Memo LEXIS 463" no-link"="" number="11" pagescheme="<span class=">1998 Tax Ct. Memo LEXIS 463">*473 Continued
| Year | 1991 | 1993 | 1994 |
| 1986 | $ 2,750 | $ 2,875 | $ 2,850 |
| 1987 | 7.250 | 7,000 | 6,900 |
| 1988 | 8,750 | 8,250 | 7,750 |
| 1989 | 11,000 | 10,250 | 9,000 |
| 1990 | 14,500 | 13,500 | 12,500 |
| 1991 | 15,000 | 14.000 | 13,250 |
| 1992 | 17,500 | 17,000 | 16,500 |
| 1993 | -- | 15,000 | 15,000 |
| 1994 | -- | -- | 16,500 |
| IBNR | 1,034 | 3,284 | 4,657 |
| Total | 77,784 | 91,159 | 104,907 |
Petitioner discounted the part of its reserves that related to occurrence-based policies during the years in issue. In those years, petitioner followed the same procedures in establishing its annual statement unpaid losses that it had used in prior years. It gave data to Tillinghast which Tillinghast used to make development method 91998 Tax Ct. Memo LEXIS 463" label="1998 Tax Ct. Memo LEXIS 463" no-link"="" number="12" pagescheme="<span class=">1998 Tax Ct. Memo LEXIS 463">*474 and pure premium method 10 projections. In the property and casualty industry, "development" is the actual experience (both paid and unpaid) regarding a loss estimate over time.
The number of petitioner's claims greatly increased in 1990 and remained at a higher level for 1991 and 1992. The severity (i.e., the average cost per claim) of petitioner's claims also increased in 1991 and 1992.
Tillinghast's 1991 and 1992 loss reserve reviews used a range bounded by a high and low end estimate of projected ultimate losses. The bounds of Tillinghast's range are the sums of the high and low end estimates of ultimate loss for each coverage year, at the December 31 valuation date. Tillinghast's ranges were relatively large because, in Hurley's opinion, medical malpractice losses are difficult to project accurately.
Tillinghast projected that, as of December 31, 1991, petitioner had ultimate losses ranging from $ 88,483,000 to $ 99,645,000 (before discounting), and that reserves ranging from $ 45,426,000 to $ 57,289,000 (before discounting) for coverage years 1978 to 1991 would 1998 Tax Ct. Memo LEXIS 463" label="1998 Tax Ct. Memo LEXIS 463" no-link"="" number="13" pagescheme="<span class=">1998 Tax Ct. Memo LEXIS 463">*475 be reasonable. As of December 31, 1992, Tillinghast projected that petitioner had ultimate losses ranging from $ 100,101,000 to $ 112,204,000 (before discounting), and that reserves ranging from $ 49,066,000 to $ 61,948,000 (before discounting) for the coverage years 1978 to 1992 would be reasonable. Tillinghast separated the projected ultimate losses by coverage year.
Joseph Perry, petitioner's vice president of finance/chief financial officer, subtracted from these ultimate loss estimates petitioner's paid losses as of December 31, 1991, and December 31, 1992, to determine a range of unpaid losses for all of the coverage years included in the 1991 and 1992 unpaid loss reserves. Petitioner reported on its annual statements that it had undiscounted unpaid losses of $ 56,847,261 for 1991 and $ 61,971,100 for 1992. Petitioner had a conservative reserve philosophy to ensure that it could pay future losses. Petitioner selected reserves below the low end of Tillinghast's range for 1986 and at the high end of Tillinghast's ranges for 1987 to 1992.
UDI examined petitioner's 1990-93 annual statements. It did not adjust the amount of unpaid losses and loss adjustment expenses that petitioner 1998 Tax Ct. Memo LEXIS 463" label="1998 Tax Ct. Memo LEXIS 463" no-link"="" number="14" pagescheme="<span class=">1998 Tax Ct. Memo LEXIS 463">*476 reported.
Estimates of unpaid losses by the medical malpractice insurance industry (both physician-owned and commercial, multiline carriers) and petitioner were similar as shown below:
| ESTIMATES OF UNPAID LOSSES FOR 1992 AND EARLIER | ||||
| As of | Medical Malpractice | Percentage | Percentage | |
| Dec. 31 | insurance industry | of original | Petitioner | of original |
| 1992 | $ 21,879,689,000 | -- | $ 61,971,100 | -- |
| 1995 | 17,709,112,000 | 81 | 48,931,000 | 79 |
1. A.M. BEST RATINGS
A.M. Best (Best) rates the financial condition of property and casualty insurers each year. From 1984 to 1992, Best gave petitioner the following ratings: 111998 Tax Ct. Memo LEXIS 463" label="1998 Tax Ct. Memo LEXIS 463" no-link"="" number="15" pagescheme="<span class=">1998 Tax Ct. Memo LEXIS 463">*477
| Year | Rating |
| 1984 | B |
| 1985 | unknown |
| 1986 | NA-7 |
| 1987 | NA |
| 1988 | B+ |
| 1989 | B+ |
| 1990 | A- |
| 1991 | A- |
| 1992 | A- |
2. SURPLUSES
Petitioner reported on its 1985 to 1992 annual statements that it had the following surpluses:
| Surplus | |
| Year | Reported |
| 1985 | $ 4,533,310 |
| 1986 | 5,481,798 |
| 1987 | 8,842,442 |
| 1988 | 10,371,232 |
| 1989 | 12,319,227 |
| 1990 | 14,382,840 |
| 1991 | 15,876,858 |
| 1992 | 18,195,874 |
3. ULTIMATE LOSS ESTIMATES
Petitioner reported the following initial estimates and 1998 Tax Ct. Memo LEXIS 463" label="1998 Tax Ct. Memo LEXIS 463" no-link"="" number="16" pagescheme="<span class=">1998 Tax Ct. Memo LEXIS 463">*478 reestimates of its ultimate losses on its annual statements from 1986 to 1996:
| Ultimate Net Losses & Allocated Loss Adjustment Expenses | ||||||
| 1 (Schedule P -- Part 2 -- Summary of Annual Statement) | ||||||
| (in thousands) | ||||||
| Year | 1986 | 1987 | 1988 | 1989 | 1990 | 1991 |
| 1986 | $ 3,865 | $ 2,492 | $ 2,414 | $ 2,191 | $ 2,191 | $ 2,205 |
| 1987 | -- | 9,077 | 7,270 | 6,655 | 7,050 | 7,250 |
| 1988 | -- | -- | 15,127 | 11,372 | 9,400 | 9,250 |
| 1989 | -- | -- | -- | 15,559 | 12,250 | 12,250 |
| 1990 | -- | -- | -- | -- | 17,403 | 15,278 |
| 1991 | -- | -- | -- | -- | -- | 17,125 |
| 1992 | -- | -- | -- | -- | -- | -- |
| 1993 | -- | -- | -- | -- | -- | -- |
| 1994 | -- | -- | -- | -- | -- | -- |
| 1995 | -- | -- | -- | -- | -- | -- |
| 1996 | -- | -- | -- | -- | -- | -- |
| Total | 3,865 | 11,569 | 24,811 | 35,777 | 48,294 | 63,358 |
Table Continued
| Ultimate Net Losses & Allocated Loss Adjustment Expenses | |||||
| 21998 Tax Ct. Memo LEXIS 463" label="1998 Tax Ct. Memo LEXIS 463" no-link"="" number="17" pagescheme="<span class=">1998 Tax Ct. Memo LEXIS 463">*479 (Schedule P -- Part 2 -- Summary of Annual Statement) | |||||
| (in thousands) | |||||
| Year | 1992 | 1993 | 1994 | 1995 | 1996 |
| 1986 | $ 2,205 | $ 2,257 | $ 2,256 | $ 2,255 | n2 |
| 1987 | 7,250 | 7,054 | 6,961 | 6,958 | $ 6,908 |
| 1988 | 9,250 | 8,624 | 8,045 | 7,751 | 7,501 |
| 1989 | 11,250 | 10,346 | 9,077 | 8,564 | 7,439 |
| 1990 | 14,528 | 13,882 | 12,905 | 12,153 | 11,129 |
| 1991 | 15,000 | 14,520 | 13,747 | 13,556 | 11,979 |
| 1992 | 17,534 | 17,607 | 17,046 | 16,769 | 16,682 |
| 1993 | -- | 16,225 | 16,110 | 15,746 | 15,797 |
| 1994 | -- | -- | 18,141 | 17,532 | 17,122 |
| 1995 | -- | -- | -- | 18,758 | 19,137 |
| 1996 | -- | -- | -- | -- | 21,520 |
| Total | 77,017 | 90,515 | 104,288 | 120,042 | 135,214 |
In 1991 and 1992, medical malpractice loss experience was favorable for coverage years 1986-90.
4. DIVIDENDS
Petitioner declared and paid dividends to its policyholders from 1988 to 1992 as follows:
| Dividend | |
| Year | amount |
| 1988 | $ 1,500,000 |
| 1989 | 1,500,000 |
| 1990 | 2,000,000 |
| 1991 | 2,000,000 |
| 1992 | 2,000,000 |
5. PREMIUMS
The premiums petitioner charged for its medical malpractice insurance policies changed as follows:
| Year | Percentage change from |
| ending | preceding year or period |
| 12/1/79 | -5.0% |
| 12/1/80 | 12.1 |
| 12/1/81 | 11.8 |
| 12/1/82 | 24.6 |
| 12/1/83 | 14.9 |
| 12/1/84 | 42.0 |
| 12/1/85 | 67.0 |
| 12/1/86 | 30.5 |
| 12/1/87 | 4.5 |
| 1/1/89 | 2.6 |
| 1/1/90 | 0.0 |
| 1/1/91 | -3.2 |
| 1/1/92 | -13.6 |
| 7/1/92 | -7.0 |
During the years at issue, petitioner bought reinsurance 12 coverage for losses falling in a certain loss "layer"; i.e., for losses and allocated loss adjustment expenses in excess of a certain minimum and below a certain maximum per loss. Petitioner's reinsurance treaty 13 during the years at issue covered losses from $ 300,000 plus an indexed amount to a maximum of $ 1 million per loss. The indexed amount equaled the product of (a) $ 25,000, and (b) the number of "December 31sts" occurring between the loss event 1998 Tax Ct. Memo LEXIS 463" label="1998 Tax Ct. Memo LEXIS 463" no-link"="" number="18" pagescheme="<span class=">1998 Tax Ct. Memo LEXIS 463">*480 and the time when petitioner first became obligated to make a payment in respect to the loss event. 14 Losses exceeding $ 1 million were covered by a separate reinsurance treaty.
Based on its pattern for paying significant claims, petitioner's average "retention" 15 by the time those claims would 1998 Tax Ct. Memo LEXIS 463" label="1998 Tax Ct. Memo LEXIS 463" no-link"="" number="19" pagescheme="<span class=">1998 Tax Ct. Memo LEXIS 463">*481 be paid was about $ 375,000 to $ 400,000 during the years in issue ($ 300,000 plus $ 25,000 for each December that passed between the coverage year and the year of final payment of the claim).
Petitioner timely filed Forms 1120-PC, U.S. Property and Casualty Insurance Company Income Tax Return, for 1991 and 1992.
Petitioner had undiscounted unpaid losses of $ 56,847,261 for 1991 and $ 61,971,100 for 1992. Petitioner reported on its Federal income tax returns that it had discounted unpaid losses of $ 45,650,249 for 1991 and $ 49,418,509 for 1992, which it deducted as part of losses incurred under
The sole issue for decision is whether petitioner may deduct $ 45,650,249 for its reserves for unpaid losses and loss adjustment expenses for 1991 and $ 49,418,509 for 1992.
On March 13, 1996, respondent sent a notice of deficiency to petitioner in which respondent determined that petitioner overstated its discounted unpaid losses by $ 5,816,776 1998 Tax Ct. Memo LEXIS 463" label="1998 Tax Ct. Memo LEXIS 463" no-link"="" number="20" pagescheme="<span class=">1998 Tax Ct. Memo LEXIS 463">*482 for 1991 and $ 3,904,930 for 1992, and that petitioner's discounted unpaid losses should have been $ 39,833,473 ($ 45,650,249 - $ 5,816,776) for 1991 and $ 39,696,803 ($ 49,418,509 - ($ 3,904,930 and $ 5,816,776)) for 1992. Respondent amended its answer after trial to assert that petitioner overstated its undiscounted unpaid losses by $ 13,070,000 for 1991 and by $ 19,394,000 for 1992, and that petitioner's undiscounted unpaid losses should have been $ 43,765,000 for 1991 and $ 42,577,000 for 1992. 161998 Tax Ct. Memo LEXIS 463" label="1998 Tax Ct. Memo LEXIS 463" no-link"="" number="21" pagescheme="<span class=">1998 Tax Ct. Memo LEXIS 463">*483 Petitioner bears the burden of proving that respondent's determination in the notice of deficiency is erroneous,
Insurance companies may deduct ordinary and necessary expenses and losses incurred.
The reserve for unpaid losses is an estimate, made at the close of a taxable year, of the insurer's liability for claims that it will be required to pay in future years.
1. EXPERT TESTIMONY
Both parties called expert witnesses to give their opinions about the reasonableness of petitioner's reserves for unpaid losses for 1991 and 1992. We may accept or reject expert testimony according to our own judgment, and we may be selective in deciding what parts of an expert's 1998 Tax Ct. Memo LEXIS 463" label="1998 Tax Ct. Memo LEXIS 463" no-link"="" number="25" pagescheme="<span class=">1998 Tax Ct. Memo LEXIS 463">*487 opinion, if any, we will accept.
There were six expert witnesses at the trial. Four were actuaries: Hurley (an actuary for Tillinghast) and Owen Gleeson (Gleeson) for petitioner, and Frederick Kilbourne (Kilbourne) and Raymond Nichols (Nichols) 20 for respondent. James Schacht (Schacht) and Lawrence Smarr (Smarr) also testified for petitioner.
2. HURLEY
We find Hurley's estimates of petitioner's reserves for unpaid losses to be reasonable. Hurley was petitioner's actuary during the years in issue. See
Hurley prepared semiannual loss reserve reviews and annual rate indication studies for petitioner. He used consistent actuarial methods and standard actuarial loss development and pure premium methods to estimate petitioner's unpaid loss reserves for 1991 and 1992.
Hurley used petitioner's and industry data in his projections. over time, he increased the weight he gave to petitioner's data relative to industry data because more of petitioner's data was available.
Hurley estimated only actual unpaid losses in establishing petitioner's annual statement unpaid losses. He based his projections on petitioner's database containing information about its past loss payments and case reserves. Hurley's report for the 1991 and 1992 annual statements estimated unpaid losses within an actuarially reasonable range. Each point in Hurley's range was reasonable. His reports met all relevant actuarial standards.
Hurley applied one exposure (i.e., pure premium) and four development methods. The development 1998 Tax Ct. Memo LEXIS 463" label="1998 Tax Ct. Memo LEXIS 463" no-link"="" number="27" pagescheme="<span class=">1998 Tax Ct. Memo LEXIS 463">*489 and exposure methods produced ultimates, which were weighted, as coverage years aged, against petitioner's loss experience reflected primarily in the development methods. Hurley's weighting of the two types of
methods was similar to a Bornhuetter-Ferguson method, 21 which is widely used for long-tailed lines of insurance like medical malpractice.
Estimates as of December 1990 of petitioner's ultimates for coverage years 1986 through 1990 were lower than estimates made previously for those years. Hurley reduced his estimate of petitioner's ultimates for 1991 and 1992 because his projected ultimates for prior coverage years were reduced.
Hurley's range was large because: (a) petitioner is a relatively modestly capitalized, single-line insurer that serves a limited geographic area; (b) it has relatively few claims, but the average cost of a claim is high; and (c) medical malpractice insurance is highly risky and longer-tailed. These facts makes sic projecting losses difficult. See
Gleeson 1998 Tax Ct. Memo LEXIS 463" label="1998 Tax Ct. Memo LEXIS 463" no-link"="" number="28" pagescheme="<span class=">1998 Tax Ct. Memo LEXIS 463">*490 and Schacht each said that Hurley's estimates of the reserves were reasonable. We find their analysis to be credible.
3. KILBOURNE
Kilbourne analyzed Hurley's loss reserve reviews and the reasonableness of petitioner's reserves for unpaid losses as shown on its 1991 and 1992 annual statements.
Kilbourne averaged the four 1991 and 1992 yearend point estimate results of Hurley's development methods with Hurley's pure premium range, based on the facts existing at the end of 1991 and 1992. Kilbourne noted that the results of the four development methods were clustered and that Hurley's reports did not indicate that any method was preferable. Kilbourne concluded that reasonable best estimates of petitioner's unpaid losses and loss adjustment expenses at the end of 1991 and 1992 could be made by averaging the four point estimate results of Hurley's four development methods with his pure premium method range.
Kilbourne analyzed Hurley's loss reserve reviews as of the end of 1991 and 1992. He concluded that Hurley's lookback approach for selecting an initial estimate of ultimate loss for the current year was flawed. He believed that Hurley's approach essentially ignored the point estimate results 1998 Tax Ct. Memo LEXIS 463" label="1998 Tax Ct. Memo LEXIS 463" no-link"="" number="29" pagescheme="<span class=">1998 Tax Ct. Memo LEXIS 463">*491 and that Hurley's ranges overstated petitioner's actuarially supported reserves for 1991 and 1992. He concluded that the best estimates of petitioner's ultimate losses were $ 43,765,000 for 1991 and $ 42,577,000 for 1992.
1. WHETHER PETITIONER'S RESERVES WERE FAIR AND REASONABLE
Respondent contends that respondent's adjustments to the unpaid loss reserves are needed to make the reserves fair and reasonable as required by
A taxpayer's reserve for unpaid losses must be fair and reasonable based on the facts in each case and the company's experience with similar cases.
Petitioner could not offset reserve deficits with reserve surpluses in another line of insurance because it wrote a single, relatively volatile line of business in a limited market. The inability to offset deficits with surpluses makes petitioner's business more risky and reasonably led petitioner to establish higher reserves.
For the years in issue, the medical malpractice industry overstated reserves to virtually 1998 Tax Ct. Memo LEXIS 463" label="1998 Tax Ct. Memo LEXIS 463" no-link"="" number="30" pagescheme="<span class=">1998 Tax Ct. Memo LEXIS 463">*492 the same extent as petitioner. This suggests that petitioner's estimates were fair and reasonable.
UDI made a triennial examination of petitioner's 1990-93 annual statements. It did not adjust petitioner's reporting of unpaid losses and loss adjustment expenses. Cf.
Hurley adjusted petitioner's loss reserves each year to account for petitioner's actual loss experience. This suggests petitioner's loss estimates were fair and reasonable. See
Hurley and Schacht testified that insurance companies have an incentive not to overstate their unpaid losses because overstating their losses may result in higher premiums, may make them less competitive with other companies, and could diminish their surplus to the point that they cannot write new policies. Petitioner's insureds would prefer 1998 Tax Ct. Memo LEXIS 463" label="1998 Tax Ct. Memo LEXIS 463" no-link"="" number="31" pagescheme="<span class=">1998 Tax Ct. Memo LEXIS 463">*493 to keep their medical malpractice insurance premiums low. This tension between petitioner and its insureds suggests that petitioner's reserve for unpaid losses was fair and reasonable.
2. WHETHER PETITIONER'S RESERVE ESTIMATES WERE WITHIN THE RANGE OF HURLEY'S ESTIMATES
Petitioner contends that its unpaid loss reserves for years ended 1991 and 1992 fall within the range of estimates made by Hurley and are fair and reasonable. Respondent erroneously contended that petitioner's reserves were above the high end of Hurley's ranges because respondent considered the discounted total reserves column rather than the undiscounted total reserves column. Petitioner selected reserves from the undiscounted column as required by sections 846(b)(1) and (2). Petitioner deducted paid losses from Tillinghast's projected reserves to account for the difference between the paid losses in petitioner's records and the paid losses in Tillinghast's records. Finally, Oslowski testified credibly that petitioner chose reserves from the high end, but not above the high end, of Tillinghast's reserve estimates. We find that petitioner's reserves were within the ranges of Tillinghast's reserve estimates.
3. WHETHER 1998 Tax Ct. Memo LEXIS 463" label="1998 Tax Ct. Memo LEXIS 463" no-link"="" number="32" pagescheme="<span class=">1998 Tax Ct. Memo LEXIS 463">*494 PETITIONER'S SELECTION OF TILLINGHAST'S HIGH END VALUES FOR 1991 AND 1992 WAS FAIR AND REASONABLE
Respondent argues that, because petitioner overstated its unpaid loss reserves for 1986 to 1990, petitioner's establishment of reserves using amounts at the high end of Tillinghast's range in 1991 and 1992 was not fair and reasonable. Respondent contends that the favorable development for 1986-90 was apparent when Tillinghast made the unpaid loss estimates in question, which meant petitioner's estimates were overstated.
We disagree. Petitioner properly considered the fact that the frequency and severity of its claims began to increase significantly in 1990. That fact, along with petitioner's prior history of inadequate reserves in 1980-85, makes reasonable petitioner's estimates of reserves for 1991 and 1992. The fact that petitioner's loss estimates for 1986-92 proved, with hindsight, to be higher than actual payments does not make petitioner's choice of values unreasonable. Petitioner's reserves for unpaid losses must be fair and reasonable, but are not required to be accurate based on hindsight.
4. WHETHER ONLY THE MIDPOINT OF AN ACTUARIALLY SOUND RANGE 1998 Tax Ct. Memo LEXIS 463" label="1998 Tax Ct. Memo LEXIS 463" no-link"="" number="33" pagescheme="<span class=">1998 Tax Ct. Memo LEXIS 463">*495 IS THE FAIR AND REASONABLE ESTIMATE
Respondent argues that, for tax purposes, the midpoint of an actuarially sound range, which respondent characterizes as "tax equipoise", is the only fair and reasonable estimate since it gives no tax advantage to either the taxpayer or to Treasury.
We disagree. Respondent cites no authority for the proposition that tax equipoise equates with the fair and reasonable standard. We have held in a different context that the high end of a range of reasonable values may be reasonable. See
5. WHETHER PETITIONER'S RESERVES MUST BE PROBABLE TO BE FAIR AND REASONABLE
Respondent argues that petitioner's reserves were not fair and reasonable because it is improbable that petitioner's losses will reach Tillinghast's highest estimates for 6 or 7 consecutive years. Respondent 1998 Tax Ct. Memo LEXIS 463" label="1998 Tax Ct. Memo LEXIS 463" no-link"="" number="34" pagescheme="<span class=">1998 Tax Ct. Memo LEXIS 463">*496 contends that, to be fair and reasonable, the estimates of unpaid losses selected by the taxpayer must be at least equally likely to occur as any other estimate of unpaid loss.
We disagree. The regulations require that the taxpayer make a fair and reasonable estimate of losses based on the facts of each case and on the taxpayer's experience with similar cases.
6. WHETHER PETITIONER'S LOOKBACK METHOD WAS PROPER
Respondent argues that Hurley's lookback method 22 for estimating the high end of its range ignored the most recent actual loss data and included the preceding year's overstated initial estimate in the initial high end estimate for the next year. For example, respondent contends that Hurley used his initial $ 16 million high end estimate made at the end of 1990 rather than using the most recent estimate for 1990 ($ 15.2 million), reestimated at the end of 1991.
We disagree that Hurley improperly reestimated the prior years' ultimate losses. Hurley reestimated the ultimate losses for each 1998 Tax Ct. Memo LEXIS 463" label="1998 Tax Ct. Memo LEXIS 463" no-link"="" number="35" pagescheme="<span class=">1998 Tax Ct. Memo LEXIS 463">*497 prior coverage year when he estimated reserves for 1991 and 1992. Hurley adjusted his ranges for 1991 and 1992 based on petitioner's favorable experience from 1986 to 1990. Respondent's contention that he merely included the prior year's initial estimates in the current year's high end estimate is mistaken. Instead, he compared the results of the development methods for the current year to the initial estimates for the prior year. In calculating the ultimate loss estimates for 1991 and 1992, Hurley used reestimated ultimate loss estimates for prior coverage years.
Respondent contends that because Hurley reestimated by gradually reducing the high end ultimate losses for the post-1985 coverage years as of the end of 1991 and 1992, he compounded petitioner's overstated reserves for those years by carrying the overstatements forward. We disagree that Hurley compounded petitioner's overstated reserves by his approach. As petitioner's loss experience began to improve, Hurley reduced the estimates for that coverage year.
7. KILBOURNE'S CRITICISM OF PETITIONER'S METHODS
Respondent argues that Kilbourne's method is reasonable and points out that it is very similar to petitioner's actual experience, 1998 Tax Ct. Memo LEXIS 463" label="1998 Tax Ct. Memo LEXIS 463" no-link"="" number="36" pagescheme="<span class=">1998 Tax Ct. Memo LEXIS 463">*498 as reflected in the development of petitioner's losses for coverage years 1987 through 1992, reestimated at the end of 1996.
Respondent's argument misses the mark. Although we agree that Kilbourne's method is reasonable, we need not decide whether his method is more reasonable than petitioner's method.
8. WHETHER GLEESON CONFIRMED THAT HURLEY'S RANGES WERE NOT ACTUARIALLY SOUND
Respondent argues that Gleeson confirmed Kilbourne's conclusion that Hurley's ranges were not actuarially sound. 23 we disagree. Gleeson applied a probability distribution 1998 Tax Ct. Memo LEXIS 463" label="1998 Tax Ct. Memo LEXIS 463" no-link"="" number="37" pagescheme="<span class=">1998 Tax Ct. Memo LEXIS 463">*499 test to Hurley's initial high end estimate of ultimate loss for 1992. Gleeson testified that Hurley's $ 17,500,000 high end estimate was a reasonable estimate of petitioner's ultimate losses for 1992 as of the end of 1992.
On cross-examination, respondent asked Gleeson to add coverage years 1987 to 1991 to his test. Respondent asked Gleeson to review Exhibits 86, 87, and 88, 241998 Tax Ct. Memo LEXIS 463" label="1998 Tax Ct. Memo LEXIS 463" no-link"="" number="38" pagescheme="<span class=">1998 Tax Ct. Memo LEXIS 463">*500 which purport to show that petitioner's net ultimate losses as reestimated for coverage years 1987 to 1992 fell increasingly to the right of the range moving back from 1992 to 1990 and fell outside the range for 1989 to 1987. Gleeson testified that respondent's computations in Exhibits 86, 87, and 88 were accurate. Respondent argues that Gleeson agreed that petitioner's estimates for coverage years 1987 to 1992, taken together, failed his test.
We disagree. Gleeson testified that respondent's probability distribution graphs did not change his conclusion that Hurley's work was reasonable. He pointed out that Exhibits 86, 87, and 88 (particularly Exhibit 87) used basic limits data, that is, data about claims that are paid or reserved at $ 100,000 or less. In Gleeson's opinion, basic limits data shows more rapid development than total limits data because the smaller and easier to settle claims are paid first. He also pointed out that the basic limits data contained in respondent's exhibits does not show claims that are paid or reserved at more than $ 100,000. Gleeson pointed out that petitioner's retention was about $ 350,000 and that the basic limits data does not account for the development between the $ 100,000 basic limits and petitioner's $ 350,000 retention amount. 25 Thus, Gleeson found Hurley's ranges to be actuarially 1998 Tax Ct. Memo LEXIS 463" label="1998 Tax Ct. Memo LEXIS 463" no-link"="" number="39" pagescheme="<span class=">1998 Tax Ct. Memo LEXIS 463">*501 sound.
9. WHETHER RESPONDENT'S USE OF HINDSIGHT WAS PROPER
Respondent points out that hindsight may be used in deciding whether to sustain respondent's proposed adjustments.
We conclude that petitioner's reserves for unpaid losses and loss adjustment expenses for 1991 and 1992 were fair and reasonable estimates of petitioner's actual unpaid losses.
To reflect the foregoing,
Decision will be entered under Rule 155.
Footnotes
1. An inter-insurance exchange is a mutual insurance company in which the members of a group insure each other's risks.↩
2. A loss is an injury sustained by a person who has a right to hold the insured liable for that injury. A loss is incurred when the event insured against occurs.
Ocean Accident & Guar. Corp. v. Southwestern Bell Tel. Co., 100 F.2d 441">100 F.2d 441 , 100 F.2d 441">446↩ (8th Cir. 1939).3. Unless otherwise indicated, we use the terms "loss" and "unpaid loss" to include both losses and loss adjustment expenses; that is, amounts paid to defend or settle claims.↩
4. Unless otherwise indicated, we use the term "coverage year" instead of "accident year" because this case involves both occurrence and claims-made coverage.↩
5. The anniversary date is the date when insurance coverage begins.
Modern Am. Life Ins. Co. v. Commissioner, 92 T.C. 1230">92 T.C. 1230 , 92 T.C. 1230">1232↩ (1989).6. Loss reserve reviews project an insurer's ultimate losses based on its loss data.↩
7. Tillinghast analyzed petitioner's rates (i.e., premiums), to help petitioner decide how much to charge its insureds in the upcoming year.↩
8. Case reserves are estimates made by an insurer of the unpaid loss amounts expected to be paid in connection with specific known claims.↩
1. Allocated loss adjustment expense is defined above at par. I-E.↩
9. Under the development method, a series of loss development factors (one for each "age" of coverage year, e.g., coverage year + 0 is the current year, coverage year + 1 is the preceding year, coverage year + 2 is the second preceding year) are developed based on the past experience of a given coverage year's paid or incurred losses over time. These factors are then multiplied by the paid or incurred losses as of the annual statement date for the corresponding age of coverage year.
10. The pure premium method projects expected losses by multiplying historical average losses per exposure unit; e.g., per doctor by the number of exposure units covered for the coverage year.↩
11. A.M. Best describes its ratings as follows:
a. A (Excellent). Assigned to companies which in Best's opinion have achieved excellent overall performance when compared to the norms of the property/casualty insurance industry. A rated insurers generally have shown a strong ability to meet their policyholder and other contractual obligations.
b. B+ (Very good). Assigned to companies which in Best's opinion have achieved very good overall performance when compared to the norms of the property/casualty insurance industry. B+ rated insurers generally have shown a very good ability to meet their policyholder and other contractual obligations.
c. B (Good). Assigned to companies which in Best's opinion have achieved good overall performance when compared to the norms of the property/casualty insurance industry. B rated insurers generally have shown a good ability to meet their policyholder and other contractual obligations.
d. NA (Not Assigned). Approximately 400 or 25 percent of the companies reported on in Best's Insurance Reports are not eligible for a Best's Rating (A+ to C). Companies with an NA rating are assigned to one of 10 classifications to identify why the company was not eligible for a Best's Rating.
e. NA-7. Below Minimum Standards. Assigned to a company that meets Best's minimum size and experience requirements, but does not meet the minimum standards for a Best's Rating of "C". ↩
1. The initial estimate is the first entry for each year in the vertical axis; the reestimates are shown on the horizon- tal axis.↩
2. The 1986 informtion was not individually available on the 1996 annual statement.
12. Reinsurance is an agreement between an insurer (the ceding company) and a second insurer (the reinsurer), under which the ceding company passes to the reinsurer some or all of the risks that the ceding company assumes through the direct underwriting of insurance policies. See
Trans City Life Ins. Co. v. Commissioner, 106 T.C. 274">106 T.C. 274 , 106 T.C. 274">278↩ (1996).13. A reinsurance treaty is a contract between two insurers in which the reinsurer agrees to provide coverage of risks that the primary insurer has already assumed under an insurance contract with another party. See
106 T.C. 274">Trans City Life Ins. Co. v. Commissioner, supra.↩ 14. For example, if, in May 1994, petitioner paid $ 800,000 to settle a covered 1991 loss event, the 1991-92 treaty would provide reimbursement of $ 425,000, that is, $ 800,000 minus the indexed amount of $ 375,000 ($ 300,000 plus $ 25,000 each for Dec. 31, 1991, Dec. 31, 1992, and Dec. 31, 1993).↩
15. Retention is the dollar level of risk up to which an insurance company is self-insured; i.e., is not reinsured. See
Dockery v. Commissioner, T.C. Memo 1998-114↩ .16. Respondent redetermined petitioner's discounted unpaid losses in the notice of deficiency, whereas respondent's adjustments in the amended answer were to petitioner's undiscounted unpaid losses. The taxpayer is required to report discounted unpaid losses on its income tax return.
Secs. 832(b)(5)(A)(ii) and84G ↩.17.
Sec. 832(c) provides in pertinent part as follows:SEC. 832(c) . DEDUCTIONS ALLOWED. -- In computing the taxable income of an insurance company subject to the tax imposed by section 831, there shall be allowed as deductions:(1) all ordinary and necessary expenses incurred, as provided in section 162 (relating to trade or business expenses);
* * * * * * *
(4) losses incurred, as defined in subsection (b)(5) of this section; ↩
18.
Sec. 832(b)(5)(A) provides as follows:(5) Losses Incurred. --
(A) In general. -- The term "losses incurred" means losses incurred during the taxable year on insurance contracts computed as follows:
(i) To losses paid during the taxable year, deduct salvage and reinsurance recovered during the taxable year.
(ii) To the result so obtained, add all unpaid losses on life insurance contracts plus all discounted unpaid losses (as defined in section 846) outstanding at the end of the taxable year and deduct all unpaid losses on life insurance contracts plus all discounted unpaid losses outstanding at the end of the preceding taxable year.
(iii) To the results so obtained, add estimated salvage and reinsurance recoverable as of the end of the preceding taxable year and deduct estimated salvage and reinsurance recoverable as of the end of the taxable year.
The amount of estimated salvage recoverable shall be determined on a discounted basis in accordance with procedures established by the Secretary. ↩
19.
Sec. 1.832-4(b), Income Tax Regs. , provides as follows:(b) Losses incurred. Every insurance company to which this section applies must be prepared to establish to the satisfaction of the district director that the part of the deduction for "losses incurred" which represents unpaid losses at the close of the taxable year comprises only actual unpaid losses. See Section 846 for rules relating to the determination of discounted unpaid losses. These losses must be stated in amounts which, based upon the facts in each case and the company's experience with similar cases, represent a fair and reasonable estimate of the amount the company will be required to pay. Amounts included in, or added to, the estimates of unpaid losses which, in the opinion of the district director, are in excess of a fair and reasonable estimate will be disallowed as a deduction. The district director may require any insurance company to submit such detailed information with respect to its actual experience as is deemed necessary to establish the reasonableness of the deduction for "losses incurred."
20. Nichols did not do an actuarial reserve study of petitioner.↩
21. The Bornhuetter-Ferguson method is an actuarial technique used to estimate the value of a company's reserves by subtracting its paid losses from its reserves.↩
22. Hurley looked back at prior years' losses to estimate current loss reserves.↩
23. Kilbourne said that Hurley's approach ignored the point estimate results of his four actuarial methods and that his ranges were skewed so that even the low ends of his ranges as of the end of 1991 and 1992 were higher than what in Kilbourne's opinion were actuarially reasonable "best estimates" of petitioner's ultimate losses.↩
24. Exhibit 86 contains several probability distribution graphs showing factors for 1987 to 1992. Exhibit 87 is a 1992 report showing updated development factors that had been contained in an attachment to Gleeson's report. Exhibit 88 contains loss development factors. These exhibits extended Gleeson's probability distribution test back through 1987.
25. Gleeson testified that petitioner's retention was about $ 350,000. We have found that it was $ 375,000-$ 400,000. The difference in retention amounts does not affect Gleeson's explanation.↩