"Is it live, or is it Memorex?", the commercial used to say in the days of cassette tapes and reel-to-reel recorders, long before digital recording entered our lexicon. We here consider an Opinion from the U.S. Tax Court where a taxpayer has taken the position that a certain type of insurance it sells is insurance for tax purposes (the real deal), but the Service contends that it is not really insurance but more in the way of a financial hedging contract.
To preface, the IRS in recent years has informally contended that there is a difference between "insurable risks" and "investment risks". The former typically involved some sort of peril, such as lighting hitting the barn, or somebody suing because an employee took offense at a sexual remark. The latter typically involved something for which an options or futures contract was available, such as hedging against rising interest rates.
Premiums made for insurance risks are, of course, deductible by the insured/payor in the year that they are paid. The insurance company/payee is allowed to treat those premiums under the tax rules relating to insurance companies, including the ability to deduct some significant amount of the moneys set aside as reserves against the current income of the insurance company.
By contrast, payments made for the purchase of an option or futures contact are not ordinarily deductible by the purchaser, and the seller must normally realize all the income received in the current year and without any offset by the seller for any amounts contributed to reserves.
The Service has essentially taken the position that something of this nature is either wholly one (insurance risks) or wholly the other (investment risks), and there is no overlap or middle ground.
This issue came to a head with a business called RVI Guaranty Co. Ltd., which here we will just shorten to "RVI". That company sold a product known as residual value insurance to various leasing companies, manufacturers and financial institutions.
Say you go and lease a new BMW for a three-year term from a dealership. The dealership is expecting that you will drive it for X number of miles and that there will be Y amount of normal wear and tear on the car, and therefore it would have a value after three years of $Z.
At the end of the three years, you return the car to the dealership. But you were a madman those three years, and the condition of the car is terrible when it is returned, such that the value of the car is only $Z minus $20,000.
It's that sort of risk that RVI's residual value insurance covered -- RVI would pay the $20,000 difference in the above example, i.e., the difference between the expected value of the leased asset and the actual value of the leased asset upon termination of the lease.
RVI issued its residual value policies for leases of primarily three categories: (1) leased automobiles; (2) commercial real estate; and (3) commercial equipment, which included aircraft and industrial equipment, etc. In 2006, this meant 754,523 automobiles worth $9.1 billion, some 2,097 parcels of real estate worth $2.1 billion, and 1,387,281 items of commercial equipment worth $4.9 billion. This by itself constituted significant risk distribution, but even within these broad categories were numerous subcategories -- there were 20 different classes of automobiles (passenger, sports cars, SUVs, etc.) alone.
The lease terms were not the same, but ranged from as short as one-year for automobile leases to the average 28 year lease for commercial property, and all ranges in between. Sometimes, RVI would cover whole pools of assets (think, "all 2014 BMWs leased by D Dealership") at once, which claims made based on the aggregate value of the pool at the conclusion of all leases.
The bottom line is that RVI was issuing its residual value policies to an incredibly diverse pool of insureds with themselves had incredibly diverse assets being insured, over 2 million individual items worth in excess of $16 billion -- this is a textbook example of risk distribution.
A key takeaway here is that the residual value policy did not cover normal wear and tear, since that value had already been taken into account on what the pricing of the asset should be when returned upon the termination of the lease. Instead, the residual value policy covered only extraordinary damage or (here is the rub) anything else which would cause the leased asset to have a lower value at the end of the lease, which could include such "economic" factors as rising interest rates or price deflation.
Because so many diverse assets were being covered, however, no single factor (except perhaps an asteroid destroying the Earth, and coincidentally reducing the value of the insured assets when their lease terminated) could would trigger a broad payout of the residual value policies. For instance, the loss of an anchor tenant at a leased strip mall in Hoboken would probably not affect the value of a leased Dodge Challenger in Oakland, or a leased 747 flying cargo between Anchorage and San Diego.
Yet, just as with insurance companies generally, there was some demonstrable correlation between RVI's overall lost history and the economy. For instance, in the first year of the Great Recession, generally accepted to be 2008, RVI's loss ratio (claims to premiums) reached an all-time high of 97.9%. In years where the economy was accellerating, RVI's loss ratios were much lower, reaching an all-time low of 0.2% in the recovery year of 2012.
The residual value policies typically would typically only pay out some amount that was lower than the anticipated return value of the leased asset, which essentially meant that the insured business retained the first layer of loss, however slight -- quite similar to a deductible. Because RVI expected its losses to be low based on its substantial experience, the cost of the residual value policies were correspondingly low, rarely exceeding $4 in premium for $100 in coverage (and sometimes as low as $0.50 per $100 in coverage).
If there was a claim against the policy, RVI demanded that its customer show ownership of the leased asset and that numerous other conditions were complied with. If RVI paid a claim (and it paid over $150 million on its 2006 policies through 2013), then the customer's rights against whoever damaged the leased asset were subrogated to RVI. There were other conditions under which RVI did not have to pay on the residual value policies; in other words, the policies were in every way like standard insurance policies, and in every way dissimilar to typical financial guarantee or options contracts.
Despite the Service's insistence that the residual value policies were not "insurance" for tax purposes, the state insurance regulators uniformly took the position that RVI's policies were "insurance" for all their purposes, and RVI was required to be licensed in all the states where it sold its policies. State insurance premium taxes paid by RVI were $639,764 in 2006 alone. Moreover, RVI was required to be independently audited, and every few years examined, by the state insurance regulators in accordance with established insurance company auditing standards, which themselves required that RVI comply with the strenuous actuarial accounting standards for its internal reserves, with which all insurance companies must comply. As an insurance company, RVI received A+ or similar ratings by the independent rating companies such as Fitch, Moody's and Standard & Poors.
Indeed, RVI was hardly the only insurance company that issued such residual vlaue policies, but instead such policies were commonly issued by other insurance companies such as AIG and Chubb, etc. It just happened that RVI was the leading issuer of such policies.
Nonetheless, RVI was audited by the Service, and the Service took the position that the residual value policies did not constitute "insurance" for Federal income tax purposes, regardless of what any mere state insurance commissioners might have thought. The crux of the Service's position was that the residual value policies were basically offering protection against an investment risk as opposed to an insurance risk. Or, to use my earlier analogy, the residual value policies were close to insuring against rising interest rates than they were lightning hitting the barn.
Having made this determination, the Service then ruled that RVI was not allowed to take a deduction against its income for the moneys that it placed into reserves, which meant that RVI ended up owing a deficiency in excess of $55 million for tax year 2006. RVI took it up to the U.S. Tax Court, which after a trial that included numerous expert witnesses, issued the opinion that we will next discuss.
The Court began its opinion by noting that:
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. [S]ome 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 Court then turned to tax law, and pointed out that neither terms "insurance" or "insurance contract" are defined by either the Tax Code or by Treasury Regulations; thus, as does every Court every court trying to figure out these terms in the tax context, reference would be made to the remarkably short opinion of the U.S. Supreme Court issued in 1941 (Helvering v. Le Gierse), where that Court rather blandly stated that insurance involves risk-shifting and risk-distributing.
Risk-shifting simply means that the risk of loss has actually passed from the insured to the insurer, although a corollary of risk-shifting is that the insurer must be well-capitalized enough to pay claims over and above the amount premiums that it receives, i.e., it has some "skin in the game".
Applied here, without RVI's policies, the insureds on the leased equipment would bear all the loss (except for their small retained layer of risk); but by purchase the residual value policies, they would completely transfer that risk to RVI instead. Even the Service's expert witnesses conceded that there was a significant transfer of risk to RVI, and this was borne out by RVI's high loss ratio in many years.
This bring us next to risk-distribution, which means that the insurer (here, RVI) is not making a singular bet on a singular loss to a single asset, but instead is pooling numerous of its risks so as to take advantage of the non-legal, actuarial concept of the Law of Large Numbers.
Although the Tax Court did not go into the issue in any depth, an explanation is in order. The Law of Large Numbers basically posits that the more risks that are underwritten, the more highly predictable the loss ratio will be. Think of it this way, if you flip a coin twice, it might come up heads twice (100% heads), or tails twice (100% tails), or heads and tails once each (50%-50%) -- there is a 1 in 3 chance of each of those outcomes. But if you flip the same coin one million times, then you should expect the number of heads and tails to be roughly 50%-50% every time you cycle through a million flips.
In other words, the Law of Large Numbers presupposes that there will be "Regression to the Mean" (the predictable average) the more risks that are taken. Anything can happen with just a few bets, but over the long term and with many bets, the outcome should be quite predictable -- this is what the Law of Large Numbers is all about.
Now, back to our Tax Court opinion. In this case, RVI had substantial risk distribution because it was issuing its residual value policies against so many diverse risks. Even the Service's expert witness on the topic eventually conceded that risk distribution was present.
The Service responded that, despite RVI's widespread risk distribution, there were factors that could cause a loss among all of RVI's assets, such as the economic conditions that cause RVI's high losses in 2008.
The Tax Court found this explanation to be unpersuasive:
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. [S]ome 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.
It was also important to the Court that previous Tax Court cases had held that nothing like "perfect" risk distribution be present, but only that it be "meaningful". In footnote 14, the Court shot down the argument by the Service that RVI's pooling of certain risks somehow negated risk distribution, since those pools merely aggregated assets and values.
The Court then went back the U.S. Supreme Court's 1941 holding in Helvering v. Le Gierse, and focused on the ruling there which held that, precisely because the Tax Code does not define "insurance", Congress meant the term to take on its meaning in its commonly accepted sense. This meant that the Court would look at five factors:
(1) Was the insurer organized, operated and regulated as an insurance company?
(2) Was the insurer adequately capitalized?
(3) Were the insurance policies legally valid and binding?
(4) Were premiums reasonable in relation to the risk of loss?
(5) Were both premiums paid and claims resolved as they should be?
That RVI met these factors could not seriously be challenged. RVI was organized as an insurance company and regulated as such in every jurisdiction where it conducted business. RVI was well capitalized according to its audits, examinations, and ratings by Standard & Poor's, etc. RVI's policies were legally binding, and in fact RVI had paid over $150 million on its 2006 policies through 2013. The premiums were negotiated by unrelated parties at arm's length.
In fact, the Service did not strongly contest these points, but instead relied on other arguments. One such argument was the Service's assertion that the RVI policies were different than most insurance policies in that they did not pay a claim for damage immediately (i.e., when the lightning hit the barn), but instead effectively deferred the payment of the claim until the lease term ended (when the farmer turned the burned barn back over to the lessor).
While that was an interesting argument, it got the Service nowhere, since the Court pointed out that when the claim is resolved doesn't make the damage itself any less random or fortuitous. The Court noted that insurance policies are not forced into any hard formulae, and considerable leeway exists for insurers and insureds to negotiate particular terms of policies that fit their particular needs. Thus, the Court similarly dispensed with the Service's argument that the RVI policies did not satisfy a vague "timing risk requirement", which is that nobody knows when an event giving rise to a loss will occur -- that simply wasn't changed because the claim was not resolved until the end of the lease.
This brings us to the Service's last-ditch argument, and the one that is probably of the most practically importance to those of us who practice in the area of insurance law: The Service argued that RVI's residual value policies covered investment risks as opposed to insurance risks. Indeed, there is a line of Tax Court authority which posits that "insurance risk" is involved where there is a hazard that is not an "investment risk".
Once again, the Court returned to that fountain of knowledge which is the short opinion in the 1941 case of Helvering v. Le Gierse. That case involved an 80 year-old woman who bought an annuity contract bundled with a life insurance contract from the same company and then died a year later, and the U.S. Supreme Court sustained a challenge that the life insurance proceeds were exempt from estate tax under the theory, essentially, that the annuity and the life insurance offset each other's risks so that there was no true insurance that was provided to her.
There was actually a long-forgotten companion case that was decided by the U.S. Supreme Court at the same time, Keller v. Comm'r., that found no insurance risk to be present where an annuity was structured so that the only real risk to the insurance company was the "risk of computational error". The Tax Court in the RVI case mentioned Keller as well.
The Le Gierse and Keller cases provided a bright contrast to the facts of the RVI case which was before the Tax Court. Because RVI was only charging premiums which averaged less than 4% of the insured value of the leased assets, RVI had considerable exposure to claims in excess of the premiums charged, and the only way that RVI could mitigate these risks was to make sure that they were underwriting huge numbers of assets, i.e., relying upon the Law of Large Numbers and a Regression to the Mean of claims such that overall RVI's policies would be profitable, even if many individual policies were not.
This is "insurance" any way you slice it, but the Service tried a different tact by arguing that RVI's policyholders were not really purchasing the residual value to protect against damages to the leased equipment through excessive wear, etc., but were really just buying the policies as a hedge against economic conditions that could drive the value of the leased assets down upon termination of the leases. Thus, the Service concluded that the risks that RVI was underwriting were really in the way of "investment risks" as opposed to "insurance risks".
A fly in the Service's ointment on this point is that, whatever position that the Service took on "investment risk", the state insurance regulators have considered policies that protect against declines in market value to be in the way of "insurance risks", and thus subject to insurance regulation. Both New York and Connecticut defined residual value policies as "insurance" by statute, and several state court cases had likewise found residual value policies to constitute insurance for state regulatory purposes.
Of course, the position that the states take on residual value policies does not ipso facto control the tax treatment under the Tax Code, but it was quite telling to the Tax Court.
The Service further argued, based on the testimony of one of its experts, that there was a difference between "insurance risk" and "pure risk". In this scenario, "insurance risk" is something caused by some peril (again, lightning hitting the barn), while "pure risk" is the risk that any factor might cause a decline in the value of the insured asset.
The Court found the Service's expert to be "unpersuasive" (largely because her testimony was not supported by the learned treatises that she cited, apparently mostly being undergraduate business law texts of a very general nature). The Court also rejected the "pure risk" and "insurance risk" distinction, citing to several examples of policies that covered "pure risk" having been held to be insurance for Federal income tax purposes, such as quite-similar mortgage guarantee insurance and municipal bond insurance. Thus, the Court:
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 [the Service's expert] cites describe municipal bond and mortgage guaranty insurance as covering "speculative risks," even though respondent insists that the triggering event is a "pure risk." [] Aristotle noted that there are at least four distinct senses in which one thing may be said to "cause" another. [] 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.
[Internal citations omitted]
We are now at the Service's last ditch, which is its argument to the effect that because RVI's insureds could have instead purchased financial products such as put options to protect against their risk of loss, the residual value policies did not constitute "insurance" in the tax sense but were instead in the way of such financial vehicles. In the Service's view, a policy may be classified as either/or, and overlap between a financial insurance and an insurance policy for Federal income tax purposes is impermissible.
That last ditch failed too, in no small part because the Service conceded that the residual value policies are not options for Federal income tax purposes. Plus, the Court noted:
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.
Moreover, the Service's position ran afoul of prior opinions which concluded that a policy could be "insurance" for Federal tax purposes even if there were similar and competing non-insurance products that existed in the financial markets.
This last proved to be the final stake driven into the Service's challenges, and it having lost on all points, the U.S. Tax Court entered judgment in favor of RVI.
ANALYSIS
Why the Service decided to bring this case is a great mystery, since on these facts it was not particularly difficult to predict that the Service would lose. We can do little more here than to presume that somebody at the Service was stuck on the idea that residual value policies were not insurance, but investment contracts, and they couldn't be un-stuck other than to be proven wrong in court. At any rate, we are left with a decisions that is very favorable for taxpayers on these facts.
The key to the result here, like so many cases involving insurance taxation, is that the numbers were big enough that the non-legal, actuarial Law of Large Numbers clearly came into play. Spreading risks over many non-affiliated insureds such that a firm Regression to the Mean can be predicted is the very essence of insurance; but that doesn't mean that insuring just a few such risks will lead to the same result.
It is important, however, not to get carried away with the results of this case. RVI was a big insurance company, covering billions of dollars in assets and receiving hundreds of millions in premiums. Whether a much smaller company (think, a smallish captive insurance company) that conducts only a small percentage of such business could likewise prevail is subject to serious question, and should not find much comfort in this opinion -- it is not negative to such companies, but it is not particularly positive either.
Nonetheless, for companies that have as a significant business the leasing of assets, this opinion answers the important question as to whether residual value policies may constitute "insurance" in the income tax sense (they certainly can, assuming other criteria are met).
Beyond that, this opinion again invalidates the Service's position that something is either "insurance" or an "investment contract", and there can be no overlap between those two categories. It also clearly invalidates the Service's contention that products which protect against risks that have significant economic variables are ipso facto investment products and not insurance products; as shown here, much deeper analysis is required.
The truth is that even prior to Le Gierse there was great confusion about what constitutes "insurance" in the Federal income tax sense, owing to the failure of Congress (and, significantly, Treasury) to define the term "insurance" in the Tax Code, and Le Gierse itself was little more than a judicial patch on that wound. With the modern expansion of the insurance marketplace into products like residual value policies which clearly have significant purely-economic influences, and the unprecedented rise of the captive insurance marketplace to cover risks that often were not traditionally insured, perhaps it is time for better legislative and administrative guidance on the subject.
The upcoming 75th birthday of Le Gierse would seem to be as propitious a time as any.
CITE AS
RVI Guaranty Co. v. Comm., 145 T.C. 9, Dkt# 27319-12 (Sep. 21, 2015).
This article at http://onforb.es/1OBBsZ1 and http://goo.gl/kjaFeA
