Summary
In R.V.I. Guar. Co. v. Commissioner, 145 T.C. at 231, we found that policies were valid and binding when the insured filed claims for covered losses and the captive insurance company paid them.
Summary of this case from Syzygy Ins. Co. v. Comm'rOpinion
145 T.C. No. 9 Docket No. 27319-12.
09-21-2015
R.V.I. GUARANTY CO., LTD. & SUBSIDIARIES, Petitioner v. COMMISSIONER OF INTERNAL REVENUE, Respondent
Dennis L. Allen, M. Kristan Rizzolo, and Daniel H. Schlueter, for petitioner. Laurie A. Nasky and John Anthony Guarnieri, for respondent.
P sold contracts for which the company's name is an acronym--"residual value insurance." The parties insured under these contracts included leasing companies, manufacturers, and financial institutions. The assets insured included passenger vehicles, commercial real estate, and commercial equipment. The insured parties were the lessors of these assets or provided financing for such leases. When pricing a lease, a lessor must estimate what residual value the asset will have when it is returned to him at the end of the lease. P insured against the risk that the actual value of the asset upon termination of the lease would be significantly lower than the expected value.
R concluded that P's policies do not constitute insurance for Federal income tax purposes. This conclusion was based chiefly on a determination that the lessors were purchasing protection against an investment risk, not an insurance risk.
1. Held: The risks insured by the policies P sold cover an insurance risk.
2. Held, further, the policies P sold constitute contracts of "insurance" for Federal income tax purposes. Dennis L. Allen, M. Kristan Rizzolo, and Daniel H. Schlueter, for petitioner. Laurie A. Nasky and John Anthony Guarnieri, for respondent.
LAUBER, Judge: During 2006 petitioner R.V.I. Guaranty Co., Ltd., & Subsidiaries (RVI or petitioner) sold contracts for which the company's name is an acronym--"residual value insurance." The parties insured under these contracts included leasing companies, manufacturers, and financial institutions. The assets insured included passenger vehicles, commercial real estate, and commercial equipment. The insured parties were the lessors of these assets or provided financing for such leases.
When pricing a lease, a lessor must estimate what residual value the asset will have when it is returned to him at the end of the lease. RVI insured against the risk that the actual value of the asset upon termination of the lease would be significantly lower than the expected value. Typically, the insured value was set slightly below the expected residual value; if the asset's actual value at the end of the lease was lower than the insured value, RVI would pay the difference.
On audit, the Internal Revenue Service (IRS or respondent) concluded that the policies RVI offers do not constitute "insurance" for Federal income tax purposes. This conclusion was based chiefly on a determination that the lessors were purchasing protection against an investment risk, not an insurance risk. Concluding that petitioner was therefore not an "insurance company" entitled to compute its taxable income using the insurance accounting rules set forth in section 832, the IRS determined a deficiency of $55,197,620 for the 2006 taxable year.
All statutory references are to the Internal Revenue Code as in effect for the tax year at issue. All Rule references are to the Tax Court Rules of Practice and Procedure. We round all dollar amounts to the nearest dollar.
Petitioner timely petitioned for redetermination of this deficiency. After concessions, the sole issue for decision is whether the RVI policies constitute contracts of "insurance" for Federal income tax purposes. We hold that they do.
The parties have filed a stipulation of settled issues resolving the other two allegations of error set forth in the petition, namely, the "alternative insurance adjustments" issue described in paragraph 4.b and the "imputed interest" issue described in paragraph 4.c.
FINDINGS OF FACT
Some of the facts have been stipulated and are so found. The stipulations of facts and the attached exhibits are incorporated by this reference. At the time petitioner filed its petition, its principal place of business was in Connecticut.
R.V.I. Guaranty Co. Ltd. (RVIG) is incorporated in Bermuda. At all times since its incorporation, it has been registered and regulated as an insurance company in compliance with the requirements of the Bermuda Insurance Act of 1978. RVIG is the common parent of an affiliated group of corporations that includes R.V.I. America Insurance Company (RVIA). RVIA was incorporated in 1994 as property and casualty (P&C) insurance company. It began business in 1995 and is domiciled in Connecticut.
During 2006 RVIA engaged exclusively in the business of issuing policies of residual value insurance. RVIA reinsured with RVIG almost all of the risk represented by these policies. Bermuda law requires insurance companies to meet specified requirements governing solvency, liquidity, minimum capital, and surplus. RVIG met or exceeded all of these requirements during 2006.
In 1999 RVIG elected under section 953(d) to be treated as a domestic corporation for Federal income tax purposes. That election was in effect during 2006 and has not been revoked. RVI filed a consolidated Federal income tax return for 2006 on Form 1120-PC, U.S. Property and Casualty Insurance Company Income Tax Return, using a calendar fiscal year.
The Policies
Petitioner issued residual value insurance policies to unrelated insureds engaged in the business of leasing assets or financing asset leases. At the inception of any lease, the lessor anticipates that the leased property will depreciate during the lease term to a probable "residual value" due to normal wear and tear. Numerous factors, however, can cause property to decline in value more precipitously than expected. These factors may include excess wear and tear, as well as macro-economic events like recession, high interest rates, or price deflation. The residual value of an asset may also be adversely affected by risks to which that particular property is subject. For example, commercial real estate might drop in value because of urban blight in a particular neighborhood or the bursting of a national real estate bubble. Industrial equipment might drop in value because of technological change or local factory closings. Passenger vehicles might drop in value because of high oil prices or a shift in consumer preferences toward battery-powered cars.
To protect against such risks, the lessor or finance company could purchase a policy of residual value insurance. In recent years, such policies have been issued by numerous well-established insurance companies, including American In- ternational Group (AIG), Chubb Group of Insurance Companies, Royal Insurance Company of America, ACE Group, QBE Group, and Great American Insurance Group. During the tax period in issue, RVIA was a leading issuer of residual value insurance policies (RVI policy or policies).
Each RVI policy indemnified the insured against loss in the event that assets insured under the policy had an actual value at lease termination lower than the insured value that the policy specified for those assets. Typically, the insured value was slightly below the expected residual value. The insured thus retained the risk for the initial layer of loss (between the expected residual value and the insured value), and RVIA indemnified the insured against the remaining risk of loss (between the insured value and a lower actual residual value).
A simple example may illustrate the mechanics of a typical RVI policy. Assume that an automobile with an initial purchase price of $20,000 is leased for three years and that its expected residual value upon lease termination is $10,000. RVIA might insure that automobile for 90% of the expected residual value, yielding an insured value of $9,000. If, at lease termination, the automobile had an actual residual value of $8,500, the RVI policy would indemnify the lessor for $500, assuming the lessor satisfied all terms and conditions of coverage. The lessor would bear the $1,000 initial layer of loss.
RVI policies typically called for a single premium payable at inception of the contract. The premiums charged depended on how much risk RVIA assumed, i.e., on the magnitude of the gap between the expected residual value and the insured value. Generally speaking, petitioner expected losses on its policies to be quite low, and it priced the insurance accordingly. The policy premium rarely exceeded $4 for each $100 of insurance protection provided and (depending on the type of property) could be as low as 50 cents for each $100 of coverage.
The RVI policies included standard terminology and policy provisions typical of insurance policies generally, including the requirement of an "insurable interest" and provisions governing claims, exclusions, payment of losses, and conditions to coverage. For RVIA to have liability under a contract, the insured had to meet various conditions precedent, e.g., paying the premium, having an ownership interest in the covered property, providing written notice of a claim, and complying with the terms of endorsements regarding return conditions. Upon payment of a loss RVIA was subrogated to any rights of recovery the insured might have against third parties concerning that property.
RVIA sometimes included in its policies other limitations on loss, such as a policy deductible. By accepting terms that limited RVIA's risk of loss, an insured could often reduce its premium. For example, the insured might elect to exclude from coverage assets of volatile value, or might accept strict "return conditions" requiring the covered property to be in excellent condition at lease termination.
Certain RVI policies provided for "pooling." Under "pooling," a single policy would cover multiple assets under leases terminating within a specific period (say one year). A "loss" would be deemed to occur if the aggregate residual value of those assets was less than their aggregate insured value.
RVIA wrote three basic types of policies--"FASB," "primary," and "hybrid." An FASB policy was one under which the insured value of the covered property was set at a level to provide the lessor with enough insurance coverage to enable it to use "direct financing lease" accounting. See Statement of Financial Accounting Standards No. 13 (a lease may be classified as a "financing lease" if the present value of the lease payments and any guaranteed portion of the residual value exceeds 90% of the value of the asset). Under a "financing lease" the lessor can accelerate income into the lease's earlier years for financial accounting purposes. A primary policy was one under which the insured value was not set at a level tied to "financing lease" accounting. A hybrid policy was one under which each insured asset was subject to both primary and FASB coverage.
FASB refers to the Financial Accounting Standards Board, the organization responsible for establishing Generally Accepted Accounting Principles (GAAP) in the United States.
RVIG's business in 2006 consisted principally of reinsuring the risks represented by RVIA's policies of residual value insurance. As measured by net unearned premiums, 97.5% of RVIG's business at year end 2006 was attributable to RVIA risks. RVIG also reinsured risks under residual value policies issued by other insurance companies. Reinsurance of risks arising under other types of contracts represented less than 1% of RVIG's business.
RVIA grouped its policies into three business segments: passenger vehicles, commercial real estate, and commercial equipment. "Commercial equipment" included aircraft, industrial equipment, and rail cars. At year end 2006 RVIA had 951policies in force insuring 714 unrelated insureds. The assets covered by these policies included 754,532 passenger vehicles, 2,097 real estate properties, and 1,387,281 pieces of commercial equipment. Within each business segment, RVIA insured a wide variety of assets, i.e., many different makes and models of automobile, various kinds of buildings in diverse geographical locations, and many different types of industrial equipment. The passenger vehicles comprised 20 different types of automobile (including pickup trucks, sedans, SUVs, and sports cars) and approximately 50 different vehicle models. The commercial real estate comprised 15 different types of properties (including retail stores, warehouses, industrial buildings, office buildings, and motels) in seven different geographic regions. And the commercial equipment comprised 30 different types of equipment, including aircraft, rail cars, construction equipment, and shipping containers.
Each business segment accounted for roughly one-third of RVIA's business as measured by remaining unearned premiums at year end 2006. The terms of the leases to which the covered assets were subject varied considerably within business segments and from one segment to another. The lease terms for vehicles were typically one to five years; the lease terms for real estate were much longer, often 28 years; the lease terms for commercial equipment varied greatly. Even where assets (such as vehicles) were subject to the same lease term (such as three years), the policies insuring them could end in different years because initiated at different times.
As measured by remaining unearned premiums, the passenger vehicle segment accounted for 31.9%, the commercial real estate segment for 34.6%, and the commercial equipment segment for 33.5% of RVIA's business. Total unearned premiums for these three segments at year end 2006 were $47.5 million, $51.6 million, and $50 million, respectively. As measured by premiums earned during 2006, the passenger vehicle segment accounted for 58.8%, the commercial real estate segment for 12.6%, and the commercial equipment segment for 28.6% of RVIA's business. The percentages for the latter two segments were smaller as measured by earned premiums because the insured assets were leased for longer terms, with the result that the premium was "earned" more slowly. Total earned premiums for the three segments during 2006 were $27.6 million, $5.9 million, and $13.4 million, respectively.
The events that could cause losses under RVI policies varied considerably. Certain macro-economic events, such as recessions, high unemployment, or unexpectedly high interest rates, could affect various insured assets similarly. But many events that could cause loss were uncorrelated. For example, technological obsolescence of a type of commercial aircraft probably would not affect the value of an office building. And risks within a given business segment were often uncorrelated. For example, the loss of a major tenant in a Chicago office building likely would not affect the value of a building leased to a restaurant in New York.
In 2007 RVIA sustained a substantial loss on a policy covering an office building in El Paso, Texas, after the building's lead tenant, El Paso Natural Gas, moved to Houston. The risk causing this loss was uncorrelated with risks that could affect the value of buildings that RVIA insured in other locations, much less the value of motels and convenience stores coming off lease 15 years later.
At year end 2006 the total insured value of RVIA-insured property was roughly $9.1 billion in the passenger vehicle segment, $2.1 billion in the commercial real estate segment, and $4.9 billion in the commercial equipment segment. Divided by type of policy, the total insured value of property covered under FASB policies was about $5.0 billion, under primary policies was about $3.7 billion, and under hybrid policies was about $7.4 billion. (All amounts ignore reinsurance).
Petitioner paid significant claims under the RVI policies and incurred significant insurance losses. On an absolute dollar basis, RVIA paid more than $150 million in claims through 2013, which included more than $28 million in claims on FASB policies. An insurance company's ratio of paid losses (including related loss adjustment expenses) to earned premiums is generally called its "loss ratio." From RVIA's inception through 2006, its cumulative loss ratio was 27.7%. From RVIA's inception through the end of 2013, its cumulative loss ratio increased to about 34%. Its annual loss ratios from 2000 through 2013 were as follows:
Respondent objected to the admissibility of financial information for 2007-2013 as post-dating the tax year in issue and "irrelevant for that reason." The Court overruled this objection. A loss under an RVI policy is payable only at the end of a lease, and many of the insured assets were subject to very long leases. By definition, therefore, many RVI policies in existence in 2006 could not have come to a payout resolution, and could not possibly have had a loss, as of year end 2006. Yet many of these policies could (and did) experience significant losses upon lease termination. In order to display accurately RVIA's loss experience under the policies it held during 2006, it is necessary to consider the complete terms of these contracts.
| Year | Loss ratio |
|---|---|
| 2000 | 1.0% |
| 2001 | 0.9% |
| 2002 | 0.3% |
| 2003 | 1.3% |
| 2004 | 64.2% |
| 2005 | 48.6% |
| 2006 | 33.2% |
| 2007 | 20.4% |
| 2008 | 97.9% |
| 2009 | 11.5% |
| 2010 | 27.1% |
| 2011 | 18.1% |
| 2012 | 0.2% |
| 2013 | 30.7% |
Regulation of Petitioner
The residual value policies were treated as "insurance" for insurance regulatory purposes during 2006 by all States in which RVIA sold products, including Connecticut, New York, Pennsylvania, Ohio, Texas, Illinois, and Georgia. RVIA was required to be (and was) licensed to sell insurance in each State in which it issued policies. It was required to pay to those States insurance premium taxes, which totaled $639,764 for 2006. It was required to file with the insurance department of each State quarterly or annual "statutory financial statements." These statements were required to be prepared in accordance with "statutory accounting principles" (SAP) prescribed by the National Association of Insurance Commissioners (NAIC). RVIA was additionally required by Connecticut, its State of domicile, to meet minimum capital and surplus requirements, which it met for 2006.
Under the SAP, an insurer may not treat a contract as "insurance" in its statutory financial statements unless it assumes a significant insurance risk under the contract and faces a reasonable possibility of incurring a significant loss. See Statement of Standard Accounting Practice 62R. RVIA determined that it was required to account for its policies as insurance--and it did in fact account for those policies as insurance--in its statutory financial statements, including those it prepared for 2006. RVIA's independent auditor, BDO Seidman LLP (BDO), examined its 2006 statutory financial statements and issued an unqualified opinion that they were fairly stated in accordance with SAP.
The Statements of Standard Accounting Practice (SSAP) provide guidance for the completion of an insurer's statutory financial statements. The SSAP are issued by NAIC and published in its Accounting Practices and Procedures Manual. SSAP 62R, cited in the text, was originally drafted to establish rules of accounting for reinsurance. However, the evidence at trial established that practitioners regularly apply its principles to direct insurance as well.
The Connecticut Insurance Department examined RVIA's 2006 statutory financial statements for compliance with SAP. During this examination, an actuary from the department met with an actuary appointed by RVIA to review its annual actuarial report prepared by PricewaterhouseCoopers LLP (PwC). Follow- ing this review, the department raised no questions concerning RVIA's accounting for its residual value policies as "insurance."
RVIA was required by Connecticut, and by each other State in which it was licensed, to secure an annual actuarial report addressing the reasonableness of its reserves for unpaid losses, loss adjustment expenses, and unearned premiums. The PwC actuarial report opined that RVIA's reserves complied with Connecticut insurance law and with accepted actuarial standards. Petitioner's expert, Michael E. Angelina, explained that, if underwriting risk had not been present, "I would expect the various reports to have highlighted this issue. This has been my past experience with the large accounting firms and regulatory agencies." Kent E. Barrett, one of respondent's experts, testified that if BDO or PwC had believed that the RVI policies did not constitute "insurance" for SAP purposes, they would have had an obligation to say so. Neither BDO nor PwC raised any question on this point.
RVIG was licensed to sell insurance and reinsurance in Bermuda. It likewise accounted for the residual value policies as "insurance" in its statutory filings. RVIG's independent auditor, Arthur Morris & Co., examined its 2006 statutory financial statements and issued an unqualified opinion that they were fairly stated in accordance with Bermuda insurance law. An independent actuary reviewed its reserves for unpaid losses and loss adjustment expenses and opined that those reserves complied with Bermuda law and accepted insurance practice.
Petitioner received "insurance strength ratings" in 2006 from the major insurance rating agencies. Fitch Ratings gave petitioner an A+ strength rating. Moody's Investors Services gave it an A3 rating. Standard & Poor's Insurer Credit Report gave it an A rating.
Tax Return and Notice of Deficiency
On its 2006 consolidated return, petitioner reported its income and expenses consistently with the requirements of section 832 governing computation of "insurance company taxable income." The IRS issued a notice of deficiency disallowing petitioner's use of insurance company accounting. It determined that:
residual value insurance policies that insure against market decline are not insurance contracts for Federal income tax purposes. It is further determined that [petitioner must] calculate its annual taxable income using IRC sections 451 and 461 instead of IRC section 832; this accounting method change is required since [petitioner] no longer qualifies as an insurance company since [it] does not meet the requirements of IRC section 831(c).
During the examination, the IRS issued Technical Advice Memorandum 201149021 (Aug. 30, 2011) outlining its position concerning these issues.
During the examination, the IRS issued Technical Advice Memorandum 201149021 (Aug. 30, 2011) outlining its position concerning these issues.
Expert Testimony
Both parties offered extensive expert testimony at trial. This testimony addressed various characteristics of "insurance" as applied to petitioner's residual value policies. These characteristics include risk shifting, risk distribution, commonly accepted notions of insurance, and the presence of insurance risk.
Risk Shifting and Risk Distribution
Petitioner offered, and the Court recognized, Michael E. Angelina, executive director of the Academy of Risk Management and Insurance at St. Joseph's University, as an expert on the topics of insurance, risk management, and actuarial science. Professor Angelina opined that "insurance at its root has two fundamental attributes: risk shifting and risk distribution." He characterized the risks against which petitioner insured as "low-frequency/high severity risks," analogizing them to earthquakes, major hurricanes, and other "catastrophic risks." He explained that petitioner distributed these risks in the same way that other P&C companies distribute catastrophic risk, e.g., by "underwriting its risks to avoid over-concentration in any one segment (passenger vehicle, commercial real estate, and commercial equipment) or geographic area." He explained that petitioner also engaged in "temporal distribution" of its risks by insuring different forms of property, with lease terms of varying length, under policies terminating in different years. This "enabled RVI to avoid a 'run on the bank' scenario in extreme economic downturns."
Professor Angelina noted that "the presence of systemic risk does not mean an insurer has failed to pool risk. * * * [T]he ability to diversify risk at the more expected levels may have adverse consequences in tail scenarios where there is no ability to diversify. This was clearly evident in 2008 during the mortgage crisis as some mortgage guarantee insurers were not able to recover from their systemic failure."
Petitioner offered, and the Court recognized, Robert S. Miccolis as an expert on insurance and actuarial science, particularly in the field of mortgage guaranty insurance. Mr. Miccolis is an actuary with Deloitte Consulting LLP and president-elect of the Casualty Actuarial Society. He opined that, for approximately 98% of the RVI policies, it was "reasonably self-evident" that risk was transferred from the insured to RVIA.
Petitioner offered, and the Court recognized, Nancy L. Litwinski, a certified public accountant (C.P.A.), as an expert in insurance accounting and the application of statutory accounting principles by insurance companies and state regulators. Ms. Litwinski testified that RVIA's policies were consistently treated as insurance for SAP purposes by RVIA, by its independent auditors, and by the Connecticut Insurance Department. This treatment, she testified, was based on the determination that the residual value policies constituted "insurance" under SAP.
Respondent offered, and the Court recognized, Charles Cook, a managing director of MBA Actuaries LLC, as an expert in insurance and actuarial science. Mr. Cook compared petitioner's policies to "property catastrophe coverages such as windstorm or flood." He opined that no meaningful risk of loss was transferred from the policyholder because RVIA "did not appear to be exposed to significant loss." He opined that petitioner's risk of loss, especially under the FASB policies, was "remote." Mr. Cook agreed that petitioner did distribute risk geographically, temporally, and among diverse business segments. But he concluded that its risk distribution and diversification were less beneficial to it than is typical for insurers because the risks it assumed were more highly correlated.
In analyzing risk transfer, Mr. Cook limited his review to losses that had occurred as of year end 2006. In opining that petitioner's risk of loss was "remote," he relied on the fact that many policies had experienced no losses as of that date. On cross-examination, it was pointed out that many of these policies could not possibly have experienced a loss as of year end 2006 because losses were payable only upon lease termination and many policies still had multiple years to run. Upon review of petitioner's post-2006 experience, Mr. Cook acknowledged that many policies for which he had computed a loss ratio of "zero" actually experienced significant losses. Mr. Cook ultimately conceded these errors, acknowl- edging that his method of computing loss ratios systematically understated the true extent of petitioner's losses.
For example, for the policy that RVIA issued to U.S. Bank in 2005, for which Mr. Cook computed a loss ratio of "zero," RVIA ultimately paid more than $12 million in claims after receiving only $8 million in premiums, producing a loss ratio in excess of 150% for the first declaration period alone. Moreover, of this $12 million in losses, $8 million was attributable to the policy's FASB coverage. This contradicted Mr. Cook's assertion that RVI's risk of loss under its FASB policies was "remote."
Respondent offered, and the Court recognized, Kent E. Barrett, a C.P.A. at Veris Consulting, as an expert with respect to International Financial Reporting Standards (IFRS), GAAP, and SAP. Mr. Barrett opined that petitioner's FASB policies had only a remote chance of loss and did not transfer significant underwriting risk. In making this assessment, Mr. Barrett relied on the fact that most of the policies on RVIA's books during 2006 "had experienced no claim payments of significance as of the end of 2006."
Commonly Accepted Notions of Insurance
Professor Angelina testified that the RVI policies have all the indicia of standard insurance policies and are treated as "insurance" by State regulators and other participants in the insurance marketplace. In formal respects, the policies "have standard sections encompassing declarations, insuring agreements, definitions, exclusions, conditions, and miscellaneous provisions." The policies have standard provisions that "define the duties of an insured after loss, how losses are to be settled, and if any remediating elements need to be reflected in the final loss settlement." RVIA maintained actuarially sound insurance reserves for its policies and accounted for all transactions using proper insurance accounting.
Mr. Barrett did not dispute these points. But he opined that the RVI policies differ from typical insurance policies in certain ways. The risk against which petitioner insures is not a fortuitous "insured event," like a car accident, a hurricane, or a fire. Rather, petitioner insures against a greater-than-anticipated decline in the economic value of property over time. As a corollary of this observation, Mr. Barrett noted that RVIA does not face what he called "timing risk," namely, the uncertainty that arises under most insurance policies as to when a covered loss will occur. Under the RVI policies, a "loss" will occur (if at all) only on the last day of the policy term, a date that is known in advance. Finally, Mr. Barrett characterized as nontraditional certain contract terms that RVIA offered to its policyholders.
Insurance Risk
Professor Angelina opined that the RVI policies cover an insurance risk and not simply an investment risk. He noted that losses on RVI policies can vary from zero to the full insured value. The premium RVIA charged was typically no more than 4% of the insured value and (for certain contracts) ranged as low as 0.5% of the insured value. Professor Angelina opined that RVIA was thus subject to underwriting risk, namely, the risk that the premiums received (and income earned thereon) will be insufficient to cover claims made under the policy. This can arise from an inaccurate assessment of future risks, from an overconcentration of risks in a particular loss-exposed area, or from macro-economic or industry-specific factors wholly outside the underwriter's control.
Mr. Miccolis opined that the risks assumed by petitioner resemble the risks assumed under policies of mortgage guaranty insurance, which are generally regarded as involving "insurance risk." In both cases, the insurer assumes a significant risk of loss "arising out of a financial transaction which is caused by an unexpected decline in the value of property after coverage begins." In both cases, the loss suffered by the insurer can be caused by local conditions in specific markets or by "macro-economic conditions, such as general unemployment, interest rates, and the state of the credit markets." The insured under a mortgage guaranty contract seeks protection against a possible investment loss--namely, diminution in the value of its loan asset--but that fact does not negate the existence of "insurance risk" under such policies. Mr. Miccolis opined that the same conclusion should follow for residual value insurance.
On cross-examination, Mr. Miccolis agreed that mortgage guaranty insurance differs from residual value insurance in one respect. Under the former, the insurer's payment obligation is triggered by the homeowner's default, a fortuitous event; under the latter, the insurer's payment obligation arises because property has declined in value as of a particular time. But despite this distinction, Mr. Miccolis testified that the two types of insurance are essentially similar: in both cases, what truly drives the insurer's loss is an underlying decline in the economic value of the insured's property or collateral.
Respondent offered, and the Court recognized, Etti Baranoff, an associate professor of finance and insurance at Virginia Commonwealth University, as an expert in risk management, insurance, and financial economics. She opined that the RVI policies are not contracts of insurance because they cover "speculative risk" as opposed to "insurance risk."
According to Dr. Baranoff, the "foundation of insurance is that it is a product responding to the management of pure risk only." A transaction involves "pure risk," she testified, if the only possible outcomes, from the insured's point of view, are "loss" or "no loss." A homeowner considering the purchase of fire insurance, for example, faces the possibility of a fire (resulting in a loss) or the possibility of no fire (resulting in no loss). The homeowner cannot enjoy a gain with respect to the risk insured against.
A lessor considering the purchase of an RVI policy, by contrast, faces three possible outcomes: "loss," "neutral," or "gain." The covered assets could depre- ciate below the expected residual value (resulting in a loss); they could depreciate to the expected residual value (yielding a neutral outcome); or they could depreciate less than expected or actually appreciate (resulting in a gain). Because the uncertainties to which the insured property is subject might generate either a loss or a gain, Dr. Baranoff characterized RVI's policies as involving "speculative risk," like a stock investment, as opposed to "pure risk." She analogized the insured under an RVI policy to an investor who, desiring to hedge his bets, purchases an option allowing him to "put" his stock to another investor if the stock declines to a specified price by a specified date.
Dr. Baranoff relied in her report on textbooks that note the distinction between "pure risk" and "speculative risk." During cross-examination, petitioner's counsel pointed out that certain of the texts she cited state that speculative risks can be insured. See, e.g., George E. Rejda & Michael J. McNamara, Principles of Risk Management and Insurance 5 (12th ed. 2014) ("Some insurers will insure institutional portfolio investments and municipal bonds against loss."). In response, Dr. Baranoff reiterated her position that "pure risk" is the only possible subject of insurance, dismissing Professor Rejda's statement to the contrary as "an uncarefully written sentence." On balance, we found her testimony argumentative and unpersuasive.
Disagreeing with Dr. Baranoff, Professor Angelina opined that insurance can cover certain speculative risks. "[T]here are many financial risks," he testified, "that now are commonly insured, such as trade credit insurance, mortgage guaranty insurance, and municipal bond insurance to name a few." The Court regarded Professor Angelina as a credible witness and found his testimony helpful.
OPINION
I. Burden of Proof
The Commissioner's determinations in a notice of deficiency are generally presumed correct, and the taxpayer bears the burden of proving those determinations erroneous. Rule 142(a); Welch v. Helvering, 290 U.S. 111, 115 (1933). Petitioner does not contend that the burden of proof shifts to respondent under section 7491(a) as to any issue of fact.
II. Petitioner's Status as an "Insurance Company"
Insurance companies are subject to the corporate income tax imposed by section 11. See secs. 801(a)(1) (life insurance companies), 831(a) (other insurance companies). The taxable income of insurance companies, however, is computed under special rules. For P&C companies, those rules are set forth in section 832, captioned "Insurance Company Taxable Income." In order to match income with anticipated loss expenses, section 832 provides (among other things) that premi- ums are generally taken into income not as received but only as "earned." See Bituminous Cas. Corp. v. Commissioner, 57 T.C. 58, 77 (1971) (observing that if "premiums were to be taxed as received and the deductions allowed only as they later became fixed, the result would be to tax very large sums of money as income when in fact those amounts will never really become income because they will have to be paid out to policyholders").
To compute its taxable income under this special regime, the taxpayer must be an "insurance company." For this purpose, "the term 'insurance company' means any company more than half of the business of which during the taxable year is the issuing of insurance or annuity contracts or the reinsuring of risks underwritten by insurance companies." Sec. 816(a) (life insurance companies, cross-referenced in sec. 831(c), other insurance companies); see sec. 1.801-3(a), Income Tax Regs. ("[I]t is the character of the business actually done in the taxable year which determines whether a company is taxable as an insurance company[.]").
Neither the Internal Revenue Code nor the Treasury Regulations define the term "insurance" or "insurance contract." The meaning of these terms for Federal income tax purposes has thus been developed chiefly through a process of common-law adjudication. In the seminal case addressing this subject, the United States Supreme Court noted that "[h]istorically and commonly insurance involves risk-shifting and risk-distributing." Helvering v. Le Gierse, 312 U.S. 531, 539 (1941). In addition to requiring risk transfer and risk distribution, the courts have considered whether the transaction constitutes insurance "in its commonly accepted sense" and whether the risk transferred is an "insurance risk." E.g., Black Hills Corp. v. Commissioner, 101 T.C. 173, 182 (1993), aff'd, 73 F.3d 799 (8th Cir. 1996). These factors establish a framework for determining "the existence of insurance for Federal tax purposes." AMERCO & Subs. v. Commissioner, 96 T.C. 18, 38 (1991), aff'd, 979 F.2d 162 (9th Cir. 1992). We conclude that the RVI policies met all of these requirements during 2006 and that petitioner was therefore taxable as an "insurance company."
A. Risk Shifting
Insurance is an arrangement that must be examined from the perspective of both the insurer and the insured. Harper Group v. Commissioner, 96 T.C. 45, 57 (1991), aff'd, 979 F.2d 1341 (9th Cir. 1992). From the insured's perspective, insurance is a risk transfer device, that is, a mechanism by which the insured obtains protection from financial loss by paying the insurer a premium. Ibid.; Black Hills Corp., 101 T.C. at 182-183. By paying a premium, the insured externalizes his risk of loss by shifting that risk to the insurer.
We have no difficulty concluding that the lessors and finance companies that purchased the RVI policies transferred to petitioner a meaningful risk of loss. As Professor Angelina explained, these companies faced a significant business risk: if the values of the leased assets declined more precipitously than expected by the end of the lease term, their lease pricing formula could generate a substantial economic loss. Absent the RVI policy, the insured would bear the entire risk associated with loss-causing events. By purchasing the policy, the insured transferred to RVIA that risk of loss, to the extent of the assets' insured values. RVIA was indisputably a well-capitalized company fully capable of paying claims and absorbing the risks transferred to it. See Harper Group, 96 T.C. at 59 (finding risk transfer where the insurer "not only was financially capable of satisfying claims made against it, but it in fact paid such claims"). The RVI policies thus transferred the "impact of a potential loss" to the insurer from the insured. See Gulf Oil Corp. v. Commissioner, 89 T.C. 1010, 1036 (1987), aff'd, 914 F.2d 396 (3d Cir. 1990).
RVIA accounted for its policies under SAP. These rules forbid an insurer in its statutory financial statements to treat a contract as "insurance" unless the insurer assumes a significant risk under the contract and faces a reasonable possibility of incurring a significant loss. See SSAP 62R. By issuing an unqualified opinion that RVIA's statutory financial statements were fairly stated under SAP, its external auditor agreed that it bore a significant insurance risk. Citing insurance accounting standards, Mr. Miccolis found it "reasonably self-evident" that risk was transferred under 98% of RVIA's policies. See FASB 113; SSAP 62R; Reinsurance Attestation Supplement 20-1, Risk Transfer Testing Practice Note. The PwC actuarial report opined that RVIA's reserves complied with Connecticut insurance law and with accepted actuarial standards, and the Connecticut Insurance Department agreed with this assessment.
The Risk Transfer Testing Practice Note, cited in the text, was originally issued in the reinsurance context. However, the evidence at trial established that practitioners regularly apply its principles to direct insurance as well. Under this principles-based standard, risk transfer is characterized as "reasonably self-evident" when: (i) potential loss under the agreement is much greater than the premium, (ii) the agreement contains standardized terms and conditions typical for the type of coverage, and (iii) the agreement does not include impermissible provisions regarding reinsurance.
Respondent's experts conceded that the RVI policies did shift some risk of loss. After being recalled, Mr. Cook informed the Court of his conclusion that RVIA's real estate segment, which accounted for 34.6% of its business during 2006, did transfer sufficient risk of loss. And Mr. Barrett acknowledged that the FASB policies, which respondent views most skeptically, transferred to RVIA "some amount of risk." As he explained, this was necessarily the case because lessors purchased FASB policies in order to obtain direct financing lease accounting, which requires the lessor to shift to the insurer a sufficient level of risk with respect to the guaranteed portion of the residual value. See Statement of Financial Accounting Standards No. 13.
The thrust of respondent's position is that the RVI policies did not transfer enough risk of loss because losses were relatively unlikely to occur. This argument is unpersuasive on both theoretical and evidentiary grounds. Both parties' experts analogized the RVI policies to "catastrophic" insurance coverage, which insures against earthquakes, major hurricanes, and other low-frequency, high-severity risks. An insurer may go many years without paying an earthquake claim; this does not mean that the insurer is failing to provide "insurance." Mr. Barrett acknowledged that, under many catastrophic coverages, the odds of a loss occurring may be quite low. He was aware of no instance in which an insurance regulator had determined that the risk of loss on a policy of direct insurance was too "remote" for the product to be treated as "insurance." And respondent offers no plausible metric by which a court could make this assessment.
In opining that insufficient risk of loss was transferred to RVIA, respondent's experts relied on the fact that, as of year end 2006, many of RVIA's policies had experienced no losses. But in computing a "loss ratio" of zero for these policies, respondent's experts committed a methodological error. Whereas RVIA received all premiums at policy inception, a loss could occur only upon lease termination; many of its policies in force at year end 2006 had three, five, or 25 years to run. By definition, no loss could possibly have occurred under such policies as of year end 2006, but major losses could (and did) occur subsequently. The absence of losses prior to 2007, therefore, was not a logical basis upon which to ground an opinion that petitioner had assumed no meaningful risk of loss under these policies. Mr. Cook ultimately conceded this error, acknowledging that his method of computing loss ratios systematically understated the true extent of petitioner's losses.
RVIA's actual loss experience demonstrates that it bore a significant risk of loss. From inception through 2006, RVIA's cumulative loss ratio was about 28%; from inception through 2013, its cumulative loss ratio was about 34%. As one would expect with catastrophic-type coverage, RVIA's loss ratio in some years was extremely low. But in other years it was as high as 49%, 64%, and (during the 2008 financial crisis) 98%. On an absolute dollar basis, RVIA paid more than $150 million in claims through 2013. Even if we consider only the FASB policies, the segment on which respondent's experts focus, RVIA paid more than $28 million in claims through 2013. All in all, we conclude that the level of risk transferred to RVIA under these policies was more than sufficient to treat them as "insurance contracts" for Federal income tax purposes.
B. Risk Distribution
From the insurer's perspective, insurance is a risk-distribution device, that is, a mechanism by which the insurer pools multiple risks of multiple insureds in order to take advantage of "the law of large numbers." This statistical phenomenon is reflected in the financial world by the diversification of investment portfolios. It is embodied in the day-to-day world by the adage, "Don't put all your eggs in one basket." Clougherty Packing Co. v. Commissioner, 811 F.2d 1297, 1300 (9th Cir. 1987), aff'g 84 T.C. 948 (1985).
Many insureds who pay premiums will not incur losses. Insuring many independent risks in return for numerous premiums thus serves to distribute risk, in effect spreading a portion of the insurer's potential liability among his insureds. See Black Hills Corp., 101 T.C. at 183; Harper Group, 96 T.C. at 59; AMERCO, 96 T.C. at 40-41. Distributing risk allows the insurer to reduce the possibility that a single costly claim will exceed the amount taken in as a premium and set aside for the payment of that claim.
RVIA insured a vast array of different risk exposures. During 2006 it had 951 policies in force covering 714 different insured parties. Besides being spread among numerous unrelated insureds, its risks were distributed in at least four ways: across business segments (passenger vehicle, commercial equipment, and real estate), across asset types within each segment, across geographic locations (for real estate), and across lease duration.
RVIA's policies during 2006 covered 754,532 passenger vehicles, 2,097 individual real estate properties, and 1,387,281 commercial equipment assets. The passenger vehicles comprised 20 different types of automobile (including pickup trucks, sedans, SUVs, and sports cars) and approximately 50 different vehicle models. The commercial real estate comprised 15 different types of properties (including retail stores, warehouses, industrial buildings, office buildings, and motels) in seven different geographic regions. And the commercial equipment comprised 30 different types of equipment, including aircraft, rail cars, construction equipment, and shipping containers.
Petitioner's insured assets were also distributed across lease terms. The policies in effect during 2006 covered assets with lease terms ranging from one to 28 years. Even within the same business segment, an event (like a real estate crash) could cause losses for some insureds yet have no adverse impact on RVIA with respect to leases terminating many years later. This temporal distribution reduced petitioner's risk because it meant that the assets it insured would be exposed to different loss-causing events occurring at different times.
Respondent's expert Mr. Cook explained: "[T]he passage of time has a significant effect on depreciation rates and market effects, and the periods of time are even longer on commercial equipment and real estate. The time spread is a valuable part of the diversification. * * * It's one of the reason[s] we didn't find the diversification to be so inferior as to not be insurance. * * * [T]he temporal distribution was one of the good things we saw."
Respondent's expert Mr. Cook acknowledged that petitioner achieved pooling, diversification, and distribution of risk. His report made a very limited claim, namely, that the risk-distribution benefits petitioner enjoyed were "less than is usual for an insurer." By the end of his testimony, however, he expressly acknowledged that he was not raising the absence of risk distribution as a reason why RVIA's policies fail to qualify as "insurance."
Undeterred, respondent contends that the RVI policies do not sufficiently distribute risk because some systemic risks, like major recessions, could cause insured assets to decline in value simultaneously. Like most insurers, RVIA did face certain systemic risks, but many of the risks against which it insured were uncorrelated. Examples of risks that affected different insured assets differently include regional economic downturns, rising fuel prices, over-supply of particular assets, technological improvements, vehicle recalls, regional industrial migration, acts of terrorism, high interest rates, decreased availability of financing, and regulatory changes like restrictive building codes. Indeed, even systemic risks like major recessions were mitigated by the temporal distribution of RVIA's risks over lease terms as long as 28 years.
Many insurers face systemic risks. Mortgage guaranty insurance, municipal bond insurance, and financial guaranty insurance all provide coverage against risk of loss attributable to adverse macro-economic conditions, such as recessions, high unemployment, high interest rates, or seizing up of credit markets. As Professor Angelina noted, some mortgage guaranty insurers during 2008-2009 "were not able to recover from their systemic failure," yet respondent concedes that the product these companies offer is "insurance." RVIA adequately distributed systemic risks, as other providers of catastrophic coverage do, by spreading its risks temporally, geographically, and across asset classes.
The legal requirement for "insurance" is that there be meaningful risk distribution; perfect independence of risks is not required. See Rent-A-Center, Inc. & Subs. v. Commissioner, 142 T.C. 1, 24 (2014) ("Risk distribution occurs when an insurer pools a large enough collection of unrelated risks (i.e., risks that are generally unaffected by the same event or circumstance"); Harper Group, 96 T.C. at 55, 59-60 (finding sufficient risk distribution where insurer insured numerous unrelated insureds even though the risks "were not statistically independent * * *, but rather were highly correlated"); Gulf Oil Corp., 89 T.C. at 1025 n.9 (stating that sufficient risk distribution may exist if risks are independent "to some minimum extent"). We have no difficulty concluding, as respondent's expert Mr. Cook ultimately did, that the RVI policies accomplish sufficient risk distribution to be classified as "insurance" for Federal tax purposes.
Respondent errs in contending that the "pooling" provisions of certain RVI policies negate risk distribution. These pooling provisions simply aggregate covered exposures; they do not negate risk distribution among the covered insureds. Equally erroneous is respondent's contention that RVIA's "deferred premium" provisions negate risk distribution. The deferred portion of the premium acted like a deductible; RVIA still collected the balance of the premium, which could be used to pay claims of other insureds. In any event, as Mr. Cook noted, fewer than 24 of RVIA's 951 policies in force during 2006 provided for deferred premiums.
C. Commonly Accepted Notions of Insurance
As the Supreme Court has observed, the absence of a statutory definition of "insurance" from the Internal Revenue Code "strengthens the assumption that Congress used the word 'insurance' in its commonly accepted sense." Le Gierse, 312 U.S. at 540; see AMERCO, 96 T.C. at 38. To determine whether an arrangement constitutes insurance in its commonly accepted sense, we have considered such factors as: (1) whether the insurer is organized, operated, and regulated as an insurance company by the States in which it does business; (2) whether the insurer is adequately capitalized; (3) whether the insurance policies are valid and binding; (4) whether the premiums are reasonable in relation to the risk of loss; and (5) whether premiums are duly paid and loss claims are duly satisfied. See Harper Group, 96 T.C. at 60; Securitas Holdings, Inc. v. Commissioner, T.C. Memo. 2014-225, at *27.
The first factor has particular significance because "Congress has delegated to the states the exclusive authority (subject to exception) to regulate the business of insurance." AMERCO, 96 T.C. at 42 (citing the McCarran-Ferguson Act, 59 Stat. 33, as amended, 15 U.S.C. secs. 1011-1015 (1998)). We have repeatedly emphasized the significance of State insurance regulation in determining whether an entity should be recognized as an "insurance company." See Sears, Roebuck & Co. v. Commissioner, 96 T.C. 61, 101 (1991), aff'd in part, rev'd in part, 972 F.2d 858 (7th Cir. 1992); Harper Group, 96 T.C. at 60; AMERCO, 96 T.C. at 42; Securitas Holdings, T.C. Memo. 2014-225, at *5-6. It is undisputed that RVIA was organized, operated, and regulated as an "insurance company" by every State in which it did business, and that RVIG was organized, operated, and regulated as an "insurance company" by its country of domicile, Bermuda.
The RVI policies likewise satisfy the other factors we have deemed relevant. RVIA and RVIG met the minimum capital requirements of their respective regulators, and both were adequately capitalized. The RVI policies were valid and binding: when covered losses occurred, the insureds filed claims and RVIA paid those claims, amounting to $150 million through 2013. The premiums charged were negotiated at arm's length between RVIA and its various insureds, none of which was related to petitioner by ownership. The RVI policies took the form of insurance and contained standard provisions typical of insurance policies generally, including the requirement of an "insurable interest." See Allied Fid. Corp. v. Commissioner, 572 F.2d 1190, 1193 (7th Cir. 1978) (requiring the insured to have an "insurable interest" in the covered assets), aff'g 66 T.C. 1068 (1976).
Respondent does not seriously challenge any of these points. Rather, he argues that the RVI policies do not qualify as insurance because they differ in certain respects from insurance policies with which most people are familiar. First, he notes that RVI policies do not pay immediately upon the happening of a "fortuitous event," like a car crash, but upon a contract's reaching its termination date. But the fact that a loss must persist to the end of a lease term does not make the events that cause the loss--recessions, interest rate spikes, or bank failures--any less random or fortuitous. The payment terms of the RVI policies are dictated by the underlying business transaction: RVIA is insuring against loss under a lease, and whether a loss has occurred cannot be known until the lease ends. This feature of the RVI policies, while perhaps atypical, does not impugn their status as "insurance." See Commissioner v. Treganowan, 183 F.2d 288, 291 (2d Cir. 1950) (contract may qualify as "life insurance" even though it lacks standard features of many life insurance policies), rev'g 13 T.C. 159 (1949); G.C.M. 39,154 (September 20, 1983) ("[D]espite the fact that the surety bonds written by the taxpayer possess certain unique characteristics not shared by many other types of insurance contracts, they nevertheless, constitute 'insurance contracts' for purposes of subchapter L[.]").
Respondent's insistence that "[f]ortuity is essential for * * * risk pooling and the law of large numbers" betrays the narrow and esoteric sense in which he employs the term "fortuity." As we have explained previously, losses under RVI policies are caused by fortuitous events outside of its control. And its policies clearly do pool risks to take advantage of the law of large numbers. Indeed, as Professor Angelina explained, the fact that losses under RVI policies occur only upon lease termination actually enhances risk pooling by "enabl[ing] RVI to avoid a 'run on the bank' scenario in extreme economic downturns." Reduced to its essentials, respondent's argument is that a loss is "fortuitous" only if payment occurs immediately or shortly after the loss-causing event occurs. Respondent cites no authority for the proposition that this feature is an essential ingredient of "insurance" for State regulatory purposes or for Federal income tax purposes.
Respondent appears to base this argument, not on the testimony of his expert witnesses, but on a passage in a scholarly article published in 2003. See Edward D. Kleinbard, "Competitive Convergence in the Financial Services Markets," 81 Taxes 225, 238 (2003). We do not read Professor Kleinbard as using the term "fortuity" in the narrow sense urged by respondent.
Respondent next argues that RVI policies fail to satisfy what he calls "the timing risk requirement." Under typical casualty policies, respondent notes, "claims are triggered by an insurable event that is uncertain as to if and when it may occur." By contrast, a loss under an RVI policy will occur (if at all) at lease termination, a date that both parties know in advance.
This argument is really a different way of phrasing respondent's previous argument, and it is unpersuasive for the same reasons. RVIA is in fact subject to an array of timing risks--e.g., whether a recession, oil price rise, or other loss-causing event will occur before or after a particular lease expires. It is uncertain under RVI's policies, as under insurance policies generally, whether or when these fortuitous events will occur. The only uncertainty absent from RVI's policies is the date on which it will be determined whether a loss has occurred. But as noted previously, this is simply a function of the underlying business transaction.
Until lease termination, the lessee possesses the covered asset and makes lease payments. It is not until the property is returned to the lessor, with a value below the expected residual value, that the lessor realizes a concrete economic loss. As Mr. Miccolis explained at trial: "[T]he lessor doesn't have a loss, doesn't have a financial impact until the property is turned in at the end of the lease." Because the economic loss does not materialize until lease termination, it is neither noteworthy nor odd that RVI defers payment of claims until that time.
Municipal bond insurance operates similarly. The bond issuer may seek bankruptcy protection long before the maturity date of the covered bond, but the bond insurer does not pay immediately upon the happening of that "fortuitous event." Rather, the insurer pays for loss of interest on the covered bond only at the interest due date, and it pays for loss of principal only at the bond's scheduled maturity date. See, e.g., OppenheimerAMT-Free Muns. v. ACA Fin. Guar. Corp., 971 N.Y.S.2d 95, 97-99 (App. Div. 2013). As under RVI policies, therefore, a loss-causing event may occur at any time during the policy term, yet the insurer is obligated to pay loss claims only at specified dates that are known both to insurer and insured in advance. Despite the absence of what respondent would call "timing risk," the Internal Revenue Code provides that municipal bond insurance policies can qualify as "insurance" for Federal income tax purposes. See sec. 832(e)(6). We see no reason why residual value insurance should be treated differently.
Finally, respondent notes that some RVI policies call for nonrefundable premiums, a feature respondent regards as atypical of insurance policies generally. But this feature, like the payment terms discussed previously, is an outgrowth of the underlying business transaction. An RVI policy will pay out, if at all, only upon lease termination. In certain circumstances--for example, if inflation develops during a long-term lease--the lessor may become confident that the residual value of his leased asset will exceed its insured value at lease termination. To prevent the insured from taking a self-serving "wait and see" attitude in this setting, RVIA may rationally choose to disallow premium refunds upon midstream policy cancellations. This type of pricing decision does not preclude the RVI policies from constituting "insurance" for Federal income tax purposes.
The IRS has recognized that there are other types of insurance, such as surety insurance, for which the policy may be made noncancellable and for which the premium therefore is nonrefundable. See G.C.M. 39,154.
In sum, we find that the RVI policies give rise to insurance "in its commonly accepted sense." Le Gierse, 312 U.S. at 540. We agree with respondent that these policies have unique features, but these features correspond to, and are driven by, the characteristics and business needs of the underlying leasing trans- actions. We do not see why an insurer's tailoring its policy terms to the risks it undertakes to insure should prevent its policies from qualifying as "insurance." The arrangements between RVIA and its insureds "are characterized as insurance for essentially all nontax purposes * * * [and a] special rule for tax purposes is not justified by either statute or case law." Sears, Roebuck, 96 T.C. at 101.
D. Insurance Risk
Though we have often noted that insurance presupposes "insurance risk," our precedents shed little light on the contours of the latter term. We have said that "[i]nsurance risk is involved when an insured faces some loss-producing hazard (not an investment risk), and an insurer accepts a payment, called a premium, as consideration for agreeing to perform some act if and when that hazard occurs." Black Hills Corp., 101 T.C. at 182. Many of our prior cases involved captive insurance arrangements in which the casualty risks involved were indisputably "insurance risks." Thus, while reciting that "'[i]nsurance risk' is required" and "investment risk is insufficient,"AMERCO, 96 T.C. at 39, our precedents do not comprehensively explain how to distinguish the one from the other.
In ascertaining whether the risk covered by the RVI policies is an "insurance risk," we will examine the arrangement "from the perspective of both the insured and the insurer." Harper Group, 96 T.C. at 57. The Supreme Court undertook the former inquiry in Le Gierse, where an 80-year-old woman purchased an annuity contract bundled with a single premium life insurance policy. The insured died one month later and her estate claimed that the life insurance proceeds were exempt from estate tax under section 302 of the Revenue Act of 1926, 44 Stat. at 70. The Court sustained the IRS' challenge to that claim.
Noting that the term "insurance" was defined neither by statute nor by regulation, the Court reasoned that "the amounts must be received as the result of a transaction which involved an actual 'insurance risk' at the time the transaction was executed." Le Gierse, 312 U.S. at 538-539. Considering the annuity and life insurance contracts together, the Court found that they "wholly fail to spell out any element of insurance risk" because "annuity and insurance are opposites; in this combination the one neutralizes the risk customarily inherent in the other." Id. at 541. Because "the total consideration was prepaid and exceeded the face value of the 'insurance' policy," the only risk effectively present from the company's viewpoint "was an investment risk similar to the risk assumed by a bank." Id. at 542.
In the companion case of Keller v. Commissioner, 312 U.S. 543, 545 (1941), the Court likewise found no "insurance risk" where the only risk borne by the insurer was the risk of computational error or the "risk that the funds might not earn enough to cover profitably the annuity payable to the decedent."
In the instant case, the RVI policies clearly involved, from the insurer's perspective, an "insurance risk" rather than a financial risk of the sort assumed by a bank. As Professor Angelina explained, RVIA was at risk for "significant underwriting losses that were not related to [its] investment returns." Depending upon the occurrence of fortuitous events, RVIA's loss under a contract could vary from zero to the full insured value. Because the premium it charged was rarely more than 4% of the insured value, it was clearly exposed to underwriting risk, namely, the risk that the premiums charged would not be enough to cover claims paid. In contrast to Le Gierse, petitioner's business model depended not simply on its investment returns, but on the ability of its underwriters to price adequately the residual value risks borne by its insureds in order to derive a sufficient pool of premiums to cover the aggregate insured losses. This is the same pricing risk assumed by insurance companies generally.
Respondent nevertheless contends that the RVI policies do not involve "insurance risk" from the perspective of the insured party. In respondent's view, the lessors and finance companies purchased the RVI policies to protect themselves against investment losses, namely, greater-than-expected decline in the market value of the assets they owned and leased. Respondent analogizes this behavior to a stock investor's purchase of a put option, which enables him to "put" the stock to another investor if the stock falls below a specified price before a specified date. From the insured's standpoint, therefore, respondent asserts that the RVI policies involve no "insurance risk," but simply an "investment risk." See Black Hills Corp., 101 T.C. at 182.
We find this argument unpersuasive for several reasons. For more than 80 years, the States have regulated as "insurance" contracts that provide coverage against decline in the market values of particular assets. In 1933 the Pennsylvania Supreme Court held that an insurer's indemnification against loss from a decline in value of real estate involved an insurance risk. See Commonwealth ex rel. Schnader v. Fid. Land Value Assur. Co., 167 A. 300, 301 (Pa. 1933). The insurance company there "insure[d] against a well-known risk, to which all landowners are subject, depreciation from the price paid." Id. at 301. The company argued that it "makes contracts merely to buy real estate, and that such contracts are not insurance." Id. at 302. The court rejected this argument, holding that the company "was clearly engaged in the business of insurance" in providing its guaranty against decline in the value of property. See id. at 303. "An insurer guarantees against loss by an event that may or may not happen. The event specifically contemplated here * * * is depreciation in value of certain land below the price paid; the loss to be indemnified is the amount of that depreciation." Id. at 302.
New York and Connecticut have by statute defined residual value policies as a form of "insurance" since 1989. See N.Y. Ins. Law secs. 1102, 1113(a)(22) (McKinney 2015); Conn. Gen. Stat. Ann. sec. 38a-92a (West 2012). In 1991 the Washington Supreme Court reached the same conclusion. See Seattle-First Nat'l Bank v. Washington Ins. Guar. Ass'n, 804 P.2d 1263 (Wash. 1991). The State of Washington had established an insurance guaranty fund to compensate policyholders in the event of insolvency of an insurance company providing insurance coverage. The question was whether contracts that "compensate[d] a lessor for a drop in the market value of its leased vehicles" constituted "insurance," thus enabling policyholders to recover from this fund losses caused by the insurer's insolvency. Id. at 1269.
The court held that the residual value policies constituted "casualty insurance," which the Washington statute defined to include insurance "[a]gainst any other kind of loss * * * properly the subject of insurance." Id. at 1267-1269 (citing Washington Revenue Code Annotated section 48.11.070). By concluding that residual value policies cover a risk of loss that is "properly the subject of insurance," the Washington Supreme Court necessarily determined that such policies involve "insurance risk." Accord Wells Fargo Credit Corp. v. Arizona Prop. & Cas. Ins. Guar. Fund, 799 P.2d 908, 910 (Ariz. Ct. App. 1990) (conclud- ing that policies "guarantee[ing] that Wells Fargo would receive a fixed value for its leased autos at the termination of the lease" constituted "casualty insurance" under Arizona law).
Consistently with this State law precedent, petitioner's regulators and external auditors have uniformly concluded that its policies involve "insurance risk." RVIA was incorporated as an insurance company in Connecticut in 1994 and has been continuously licensed to conduct the business of insurance by its domicile and by all other States in which it transacts business. Because RVIA sells "insurance," it is required to pay to these States insurance premium taxes (which it has paid) and to meet minimum solvency requirements (which it has met).
During 2006 RVIA was required to file statutory financial statements prepared in accordance with SAP. These rules forbid an insurer in its statutory financial statements to treat a contract as "insurance" unless the insurer assumes a significant risk under the contract and faces a reasonable possibility of incurring a significant loss. See SSAP 62R. By issuing an unqualified opinion that RVIA's statutory financial statements were fairly stated under SAP, its external auditor agreed that it bore a significant "insurance risk." The Connecticut Insurance Department examined RVIA's 2006 statutory financial statements for compliance with SAP and agreed with this assessment. As Professor Angelina noted, if there had been no underwriting or insurance risk, "I would expect the various reports to have highlighted this issue. This has been my past experience with the large accounting firms and regulatory agencies."
As noted earlier, Congress generally has delegated to the individual States the authority to regulate the business of insurance. See McCarran-Ferguson Act, Pub. L. No. 79-15, 59 Stat. 33 (1945) (codified as amended at 15 U.S.C. secs. 1011-1015 (2006)); Barnett Bank of Marion County, N.A. v. Nelson, 517 U.S. 25, 40 (1996) ("Congress 'moved quickly,' enacting the McCarran-Ferguson Act 'to restore the supremacy of the States in the realm of insurance regulation.'") (quoting Dep't of Treasury v. Fabe, 508 U.S. 491, 500 (1993)). State courts and State regulators have consistently recognized as "insurance" residual value policies issued, not only by RVIA, but also by AIG, Chubb Group, ACE Group, Royal Insurance Company of America, and other well-established insurance companies. The uniform conclusion of State insurance regulators that the RVI policies involve "insurance risk," while "not dispositive of the issue before us, * * * [does] inform our decision." AMERCO, 96 T.C. at 42.
Against this consensus of insurance regulators, insurance auditors, and the insurance marketplace, respondent offers Dr. Baranoff's opinion that the RVI policies are not "insurance" because they do not cover a "pure risk." According to Dr. Baranoff, "pure risk" exists only in a binary situation where the only possible outcomes are "loss" or "no loss." A lessor who buys an RVI policy, she notes, has the potential to enjoy a gain on the underlying leasing transaction, e.g., if the leased assets appreciate rather than depreciate in value. In her view, the RVI policies thus protect the insured, not from an "insurance risk," but from a "speculative" or market risk. Respondent invites us to adopt this "pure risk" test as a bright-line rule to demarcate "insurance risk" from "investment risk."
While the covered assets could conceivably appreciate in value from lease inception, RVIA would never pay a claim in that event. The RVI policies indemnified the insured against economic loss if the actual residual value of the asset at lease expiration was less than its insured value. RVIA did not insure against reduction in value attributable to normal wear and tear and did not cover the initial layer of an insured's loss. In short, RVIA would pay a claim, as a fire insurance company would pay a claim, only where the insured had suffered a sizable economic loss. This is important because insurance generally acts only to indemnify the insured. See Epmeier v. United States, 199 F.2d 508 (7th Cir. 1952).
We decline this invitation. In support of her theory that "pure risk" is the only possible subject of "insurance," Dr. Baranoff relies, not on actual experience with the insurance market, but on citations from textbooks designed for college business students. While these authors note the distinction between "pure risk" and other types of risk, they do not support her contention that "pure risk" is the only possible subject of "insurance." Rather, they state (correctly) that insurance is "generally" or "normally" targeted to pure risks. In a later edition of his text, which Dr. Baranoff does not cite, Professor Rejda notes that while insurers "generally concentrate" on insuring pure risk, there are "certain exceptions." "Some insurers," he explains, "will insure institutional portfolio investments and municipal bonds against loss." Rejda & McNamara, supra, at 5. Other scholars describe the difference between "pure risks" and "speculative risks" as "[t]o a large extent * * * semantic," concluding that "[n]othing in the nature of speculative risk unequivocally precludes the writing of insurance." C. Arthur Williams, Jr., Michael L. Smith, & Peter C. Young, Risk Management and Insurance 8, 384 (8th ed. 1998).
See Mark S. Dorfman, Introduction to Risk Management and Insurance 8 (8th ed. 2005) ("Most speculative loss exposures are not subject to insurance." (Emphasis added)); Scott E. Harrington & Gregory R. Niehaus, Risk Management and Insurance 6-7 (1999) ("Insurance contracts generally are not used to * * * finance losses associated with price risks." (Emphasis added)); George E. Rejda, Principles of Risk Management and Insurance 6 (8th ed. 2003) ("[P]rivate insurers generally insure only pure risks * * * [and] speculative risks generally are not considered insurable." (Emphasis added)); Emmett Vaughan & Therese Vaughan, Fundamentals of Risk and Insurance 6-7 (11th ed. 2014) ("The distinction between pure risk and speculative risks is an important one because normally only pure risks are insurable." (Emphasis added)).
Confronted on cross-examination with the statements of insurance scholars that insurance can cover speculative risks, Dr. Baranoff insisted that those statements are actually consistent with her view that insurance covers "pure risk" only. She dismissed Professor Rejda's most recent statement to the contrary as "an uncarefully written sentence." In the end, Dr. Baranoff was unable to explain how her view lined up with those of the authors she cited, and we found her testimony unpersuasive.
During trial and in post-trial briefs, the parties and their experts extensively cited two standard treatises on insurance law, 1 Couch on Insurance 3d (2015) and The New Appleman on Insurance Law (2015). Neither of these treatises uses the term "pure risk" when defining the meaning of "insurance" at common law. Judicial precedent likewise suggests no limitation that would restrict "insurance" to the binary situation of "loss or no loss." Most cases require only that the insured shift to the insurer the risk from a "hazard," a "specific contingency," or some "direct or indirect economic loss."
See, e.g., Epmeier, 199 F.2d at 510 (noting that insurer indemnifies insured "against loss arising from certain specified contingencies or perils"); Black Hills Corp., 101 T.C. at 182 (noting that insurer agrees "to perform some act if and when * * * [a specified] hazard occurs"); AMERCO, 96 T.C. at 38 (noting that insured "faces some hazard" and insurer agrees "to perform some act if or when the loss event occurs"); Allied Fidelity Corp., 66 T.C. at 1074 (defining insurance as an agreement to protect insured "against a direct or indirect economic loss arising from a defined contingency"). A few of our cases have found as a fact that "pure risk" existed in a particular case, or have summarized expert testimony noting that insurance typically covers "pure risk." See Sears, Roebuck, 96 T.C. at 65, 92-93; AMERCO, 96 T.C. at 33-34; Humana Inc. v. Commissioner, 88 T.C. 197, 209 (1987), aff'd in part, rev'd in part, 881 F.2d 247 (6th Cir. 1989). In none of these cases were we asked to decide whether "pure risk" is the only possible subject of "insurance" for Federal income tax purposes or to determine whether a particular contract failed to qualify as "insurance" because it provided coverage for something other than a "pure risk." In AMERCO, 979 F.2d at 167, the Court of Appeals for the Ninth Circuit rebutted one of the IRS' arguments by stating that insurance risk exists where "[t]he only possible outcomes are loss or no loss." That statement was dictum because the existence of "insurance risk" was not at issue in AMERCO. The Ninth Circuit affirmed our Court's finding of fact that "there was an insurance risk involved" because "the AMERCO Group undoubtedly faced potential hazards from its operations which constituted insurable risks." Id. at 165.
Respondent's "pure risk" position lacks practical as well as theoretical support, for if we accepted his submission several familiar types of insurance would seem to be disqualified as such. As early as 1932, the Supreme Court held that mortgage guaranty insurance constitutes "insurance" for Federal income tax purposes. See United States v. Home Tit. Ins. Co., 285 U.S. 191, 195 (1932) ("The guaranty of payment of the principal and interest of mortgage loans constitutes insurance."); Bowers v. Lawyers' Mortg. Co., 285 U.S. 182, 189 (1932) ("Undoubtedly the guaranties contained in the policies and participation certificates were in legal effect contracts of insurance."). The Internal Revenue Code explicitly recognizes both "mortgage guaranty insurance" and "lease guaranty insurance" as "insurance" for Federal income tax purposes. See sec. 832(b)(1)(E), (c)(13), (e)(3) (specifying rules for computation of "insurance company taxable income" by "a company which writes mortgage guaranty insurance"); sec. 832(e)(6) (same, for "a company which writes lease guaranty insurance").
Mortgage guaranty insurance protects a mortgage lender from the risk that his collateral may decline in value and be insufficient to cover the remaining loan balance in the event of foreclosure. Residual value and mortgage guaranty insurance thus cover substantially the same risk: unexpected decline in the market value of the insured's interest in property. As Mr. Miccolis explained, both forms of insurance "provide protection against a contingent financial loss arising out of a financial transaction which is caused by an unexpected decline in the value of property after coverage begins." The two types of insurance, he noted, involve risks and risk characteristics that "are comparable with respect to substance, causation, events, conditions, and financial impact." In both cases, the value of the covered assets may be adversely affected by fortuitous events specific to the particular property as well as by macro-economic conditions such as interest rates, unemployment, inflation, deflation, and unstable credit markets. These are the same risk exposures that Dr. Baranoff cites in support of her position that the RVI policies cover an uninsurable "speculative risk."
Municipal bond insurance, which gained prominence in the United States during the 1970s, likewise provides coverage against speculative risk. See generally 120 Cong. Rec. 28114, 28115 (1974). The Internal Revenue Code explicitly recognizes municipal bond insurance as "insurance" for Federal income tax purposes. See sec. 832(e)(6) (specifying rules for computation of "insurance company taxable income" by "a company which writes * * * insurance on obligations the interest on which is excludable from gross income under section 103"). This form of insurance protects the bondholder against loss of profit on his investment by guaranteeing that he will receive payment of interest and repayment of principal if the issuer fails to pay. As in the case of residual value insurance, the insured does not face a binary situation of "loss or no loss," but has the possibility of gain or loss on his bond investment. And losses under municipal bond policies, as under mortgage guaranty and residual value policies, may be linked to macro-economic factors as well as factors specific to the particular insured asset.
Respondent insists that these two types of coverage differ from residual value coverage in terms of the "triggering event." Under a mortgage guaranty policy, for example, the insurer's payment obligation is triggered by the homeowner's default, which respondent views as a random and "fortuitous" event. Because default is "an occurrence from which the insured cannot profit," respon- dent views mortgage guaranty coverage as involving "pure risk" even though the collateral securing the loan (a home) represents an investment that can rise or fall in value. Under a residual value policy, by contrast, the event that triggers the insurer's payment obligation is simply a decline in the value of the insured property as of a certain date, namely, lease expiration.
We think respondent is confusing the events that may trigger a payment obligation with the events that actually cause the loss. The homeowner's default does not necessarily cause a loss; if the homeowner defaults because he has become unemployed, but the home is worth substantially more than the outstanding mortgage balance, the mortgagee upon foreclosure will experience no loss and will make no claim on the insurer. Under mortgage guaranty insurance, what actually causes the loss are the events responsible for the decline in the value of the house that serves as collateral for the loan. The same is true for residual value insurance.
In any event, we find respondent's attempt to distinguish between a "pure risk" and a "speculative risk" in this setting as essentially metaphysical in nature. The textbooks that Dr. Baranoff cites describe municipal bond and mortgage guaranty insurance as covering "speculative risks," even though respondent insists that the triggering event is a "pure risk." See, e.g., Rejda & McNamara, supra, at 5; S.S. Huebner, et al., Property and Liability Insurance 366-367 (4th ed. 1996) (describing municipal bond insurance as providing coverage against speculative risk). Aristotle noted that there are at least four distinct senses in which one thing may be said to "cause" another. Physics, bk. II, ch. 3. Respondent's efforts to split hairs by disentangling the causes of "loss" are philosophically interesting. But we do not think they carry much weight in determining whether the RVI policies constitute "insurance" for Federal income tax purposes.
One might describe the homeowner's default as the "but for" cause of the mortgage guarantor's loss, whereas the bursting of a national real estate bubble might be the "efficient" cause.
Finally, respondent urges that we find the RVI policies to entail mere "investment risk" by analogizing its policyholders to investors who have purchased put options to protect their stock. The problem with this argument is that the insureds are not investors and the policies are not derivative products. Investors invariably purchase stock in the hope that it will appreciate in value, enabling them to sell the shares for a capital gain. The assets petitioner insured are not investment assets; in the hands of the lessors or finance companies, they are ordinary business assets in the nature of inventory or equipment. The insureds do not acquire these assets expecting them to appreciate in value and be sold to generate gain. To the contrary, the insureds typically expect these assets to decline in value, but believe that they can nevertheless be leased profitably if the lessor's lease-pricing formula works as expected.
The insureds purchase insurance from RVIA to protect against the risk that unexpected events will wreak havoc with these lease-pricing formulas and generate an ordinary business loss instead of a profit. This is not an investment risk; it is a risk at the very heart of the lessor's business model. In comparison with typical stock investors, therefore, the insureds under the RVI policies are at the opposite end of the bell curve.
Analogizing the RVI policies to put options, moreover, is little more than a simile. In the real world, put options are typically settled for cash rather than by actual transfer of the underlying shares. At a conceptual level, many insurance products could be likened to put options. A mortgage guaranty policy, for example, could be said to give the policyholder the right to put the mortgage loan to the insurer unless the insurer pays the insured the difference between the remaining balance of the loan (the strike price) and its value on the exercise date. Even a fire insurance policy could be likened to a put on the fire-damaged house that is settled by the insurer's payment of the damage claim.
The parties agree that the RVI policies are not and cannot be taxable for Federal income tax purposes as derivative products. These policies were priced, sold, and regulated as insurance products. For financial and securities regulatory purposes, the policies cannot be treated as put options because (among other reasons) they are regulated by the States as "insurance." See 17 C.F.R. sec. 1.3(xxx)(4)(i)(A) (2014); id. sec. 240.3a69-1(a)(1).
The courts have long held that a product can be "insurance" even though competing products exist in the financial marketplace. In 1931 the Court of Appeals for the Second Circuit rejected the argument that a mortgage guaranty contract was not "insurance" because "banking corporations may also sell mortgages with their guaranty." Home Title Ins. Co., 50 F.2d 107, 111 (2d Cir. 1931) (concluding that the State's recognition and regulation of the issuer as an insurance company "should turn the scales, if the question hangs in doubt"). And in 1933 the Pennsylvania Supreme Court rejected the argument that a contract guaranteeing the value of land could not be "insurance" because it resembled "a real estate option." See Fid. Land Value Assur. Co., 167 A. at 302. When it comes to mitigating risk, there may be more than one way to skin the cat. The existence of other strategies does not mean that the strategy chosen is not "insurance" or that product purchased involves no "insurance risk."
In Chief Counsel Advisory 201511021, 2015 WL 1094778 (Mar. 13, 2015), the IRS concluded that contracts under which a captive insurer indemnified its manufacturing affiliates against "loss of earnings" attributable to foreign currency swings did not constitute "insurance" for Federal income tax purposes. The IRS noted, among other things, that the contracts "provide[d] a reasonable approximation" of the loss suffered by the affiliates, rather than "measur[ing] the actual loss suffered by the change in exchange rate." Cf. sec. 998 (providing for the tax treatment of certain foreign currency transactions). We express no view on whether these contracts would constitute "insurance" under the analysis set forth in this Opinion. --------
For all these reasons, we reject respondent's contention that the RVI policies involve an uninsurable "investment risk." These policies were designed and marketed as insurance products. Similar products were sold in the insurance market by other major insurance companies. These policies were undergirded by insurance strength ratings from the major insurance rating agencies. For more than 80 years the courts have recognized that contracts insuring against the risk that property will decline in value can involve "insurance risk." The types of events that cause losses under these policies closely resemble the events that cause losses under policies of mortgage guaranty and municipal bond insurance. Most importantly, every State in which petitioner does business recognizes these policies as involving insurance risk and regulates them as "insurance." Respondent is correct that these policies have some features that are atypical of what might be called "standard" insurance policies. But these differences are driven by the economics of the underlying business transaction and do not nullify the existence of "insurance risk."
E. Conclusion
Our analysis of insurance risk, risk transfer, risk distribution, and the commonly accepted notions of insurance convinces us that the RVI policies are "insurance contracts" for Federal income tax purposes. Because more than half of petitioner's business during 2006 consisted of issuing "insurance contracts," petitioner was for that year an "insurance company" within the meaning of section 831(c) and was required to compute its "insurance company taxable income" under section 832.
To reflect the foregoing and the parties' concessions,
Decision will be entered under Rule 155.