Medicare Supplement and FASB’s Long-Duration Targeted Improvements Project Part I: Context
By Rowen B. Bell
Health Watch, September 2021
Editor’s Note: This article is part of a two-part series. Part II is also published in this issue and can be found here.
In 2018, the Financial Accounting Standards Board (FASB) adopted significant revisions to the U.S. GAAP insurance accounting literature, via Accounting Standards Update (ASU) 2018-12. The title of ASU 2018-12 is “Targeted Improvements to the Accounting for Long-Duration Contracts,” and as a result, a common acronym for ASU 2018-12 is LDTI (long-duration targeted improvements). As of this writing, LDTI is scheduled to take effect for the calendar year 2023 GAAP financial statements issued by public companies, but not until calendar year 2025 with respect to GAAP financial statements issued by private companies.[1]
Implementation of LDTI has been a major undertaking for the life insurance industry. Awareness of LDTI in the health insurance industry is considerably less, owing primarily to the fact that most major health insurance products—including commercial medical, Medicare Advantage and Medicaid managed care—are classified as short-duration rather than long-duration insurance, meaning that they were unaffected by the new LDTI guidance. Most of the attention paid to LDTI within the U.S. health actuarial community has been on long-duration health insurance products such as long-term care (LTC), individual disability income (IDI) and critical illness, which are written predominantly by life insurers.
This two-part article focuses on the implications of LDTI for Medicare Supplement insurance. For many health insurers, Medicare Supplement may be their only material long-duration line of business, and LDTI implementation for Medicare Supplement may not yet have garnered a high level of attention. Moreover, as we discuss herein, several aspects of how Medicare Supplement insurance is priced and managed lead to somewhat unusual results when applying the LDTI guidance to this product. In particular, LDTI disturbs the conventional wisdom that only issue-age-rated Medicare Supplement policies need GAAP benefit reserves.
Part I of this article provides relevant background on both LDTI and Medicare Supplement insurance, as well as on current GAAP accounting practices for this product. In Part II, the focus shifts to the impact of LDTI on Medicare Supplement.
A Brief LDTI Refresher
Before turning to Medicare Supplement considerations, we will highlight a few aspects of the LDTI accounting model that are relevant to our discussion. This discussion is, by design, concise and targeted; the reader is encouraged to seek out other sources for a more comprehensive discussion of these and other important LDTI topics.
No Lock-in Principle
Under existing U.S. GAAP, the assumptions used in the liability for future policy benefits (which we will also refer to as “benefit reserves”) are locked in at policy issuance and are only updated later if loss recognition testing indicates that the liability is no longer adequate, in which case the former assumption basis is “unlocked” and replaced with a new locked-in set of assumptions.
Under LDTI, however, the liability assumptions are not locked in; instead they are updated over time to reflect current best estimates. Moreover, that update process is both prospective and retrospective in nature, taking into account not only changes in future assumptions but also actual past experience. In a loose sense, and using terms familiar to health actuaries, under LDTI the benefit reserves reflect the difference (if any) between the current estimate of the future loss ratio and the current estimate of the lifetime loss ratio, in both cases measured on a discounted basis.
Cohorts
Under existing U.S. GAAP, it is common for benefit reserves to be calculated on a seriatim (policy-by-policy) basis. However, under LDTI, the insurer needs to define a structure of cohorts into which policies are grouped for purposes of benefit reserve calculations. While different policies within a cohort may require different assumptions when projecting future cash flows, ultimately the calculation of the LDTI benefit reserves is performed at the cohort level rather than at the policy level. Also, there is now clear guidance stating that the reserve for each cohort is floored at zero, even in a situation where the pure mathematics of the calculation would imply a negative reserve (e.g., because the future loss ratio for the cohort is expected to be less than its lifetime loss ratio).[2]
The insurer has considerable latitude under LDTI to determine its cohort structure, but one hard-and-fast rule is that a cohort cannot contain policies issued more than a year apart.[3] As such, a common emerging approach in LDTI implementation is for an insurer first to define different blocks of business, and then to have one LDTI cohort for each combination of block and calendar year of issuance.
One Reserve Instead of Two
Under existing U.S. GAAP, the estimation of the liability for future policy benefits has historically been kept completely separate from the estimation of liabilities for claims that have already been incurred. This separation mirrors the situation under U.S. statutory accounting, where there are distinct concepts of contract reserves (ALR) versus claim reserves/liabilities (DLR/IBNP), with each subject to its own requirements.
When ASU 2018-12 was first adopted, a prevailing belief among industry participants was that the new guidance did not impact how insurers would estimate unpaid claim liabilities arising from long-duration contracts. However, as implementation guidance was crafted by the accounting profession, a new view emerged.
In 2020, the American Institute of Certified Public Accountants (AICPA) adopted guidance stating that under LDTI the liability for future policy benefits is supposed to incorporate “all payments under the contract, including future expected claims and claims for which the . . . insurance event has occurred but . . . have not yet been paid, which obviates the need for a separate claim liability.”[4] As such, under LDTI there is really only one reserve, not two.[5]
The AICPA also specifies that an insurer could, if desired, bifurcate the LDTI liability into an “incurred claims” piece versus a “future claims” piece for financial statement presentation purposes, mimicking current practice. However, in sanctioning that practice, the AICPA makes it clear that even if the claim liability piece were presented separately, the total liability on the balance sheet should be the same as if the liability were presented in a unitary fashion.[6]
Acquisition Cost Capitalization and Amortization
Under existing U.S. GAAP, to the extent an insurer’s costs for acquiring a contract (e.g., broker commissions) are incurred in a nonlevel fashion over time, then the insurer capitalizes certain of those costs, creating a deferred acquisition cost (DAC) asset, and amortizes those costs into expense over the lifetime of the product. The most straightforward way in which this occurs is with so-called heaped commission structures, where the commissions paid by the insurer in the early years that the contract is in force are higher than those paid in later years.
For most long-duration health insurance products, today the amortization of DAC is done in proportion to premiums. As a consequence, under existing GAAP DAC can be thought of, loosely speaking, as the vehicle by which the insurer expects to realize a relatively level ratio of net commission expense to premium over the lifetime of a policy, even though the cash-basis commissions do not represent a level percentage of premiums.
LDTI retains the DAC concept but with some modifications, three of which we highlight here.
First, one of the stated objectives of LDTI was to simplify the amortization of deferred acquisition costs. The principle articulated in the new guidance is that DAC be amortized “on a straight-line basis.” What exactly that means may be open to some interpretation, but there is clear guidance that “the resulting amortization amount shall not be a function of revenue or profit emergence.”[7]
Second, the DAC asset is no longer credited with interest, unlike under existing U.S. GAAP.[8]
Third, it is now made clear under LDTI that certain types of acquisition costs are not eligible for deferral, including those “that vary in a constant relationship to premiums or insurance in force” and those “that tend to be incurred in a level amount from period to period.” It is similarly made clear that “ultimate level commissions . . . shall be charged to expense” immediately.[9]
What Makes Medicare Supplement Special?
Medicare Supplement is a guaranteed-renewable health insurance product, most commonly sold to individuals at the point they become eligible for Medicare at age 65. As the name suggests, it provides coverage that supplements the benefits received by individuals enrolled in traditional Medicare, as contrasted with insurer-provided Medicare Advantage plans that substitute for traditional Medicare. Per NAIC data, in 2019 U.S. insurers wrote more than $32 billion of Medicare Supplement premiums, with 25 different insurance holding company groups each having more than $250 million.
From a U.S. GAAP standpoint, insurers have generally concluded—either explicitly or implicitly—that Medicare Supplement meets the criteria in ASC 944 to be classified as a long-duration rather than a short-duration insurance contract.[10] Having said that, there are several features of Medicare Supplement insurance that, taken together, make this product somewhat unique in the universe of long-duration insurance contracts.
Routine Premium Increases
Unlike most long-duration contracts, Medicare Supplement premiums are typically guaranteed for only 12 months at a time, and moreover there is an expectation when the contract is issued that premiums will need to increase on an annual basis. Contrast this with noncancelable IDI, where premiums are fully guaranteed, or LTC, where future premium increases are possible but comparatively less frequent and are typically not assumed in the pricing process. Fundamentally, these routine premium increases reflect the fact that health care costs (and also Medicare Part A and Part B deductibles) are expected to keep increasing over time, in light of not only general inflation but also changes in health care delivery patterns. The premium increases also reflect the pricing structure in effect, as discussed immediately below.
Different Pricing Structures
With Medicare Supplement, there is generally an increasing relationship between an individual’s age and their expected claim costs. However, broadly speaking there are three different pricing structures in effect for Medicare Supplement, with some state insurance regulators having imposed one particular structure over the others for public policy reasons:
- Attained-age-rated. Here, the premiums an individual is charged are a function of their current age. Consequently, some portion of the annual premium increases that an individual receives reflects the impact that the individual’s aging has on expected claim costs. As such, when a policy is issued, the insurer may expect that each policy’s expected loss ratio ought to remain relatively flat over time as the insured ages, since future annual premium increases are expected to keep pace with the impact of aging on claims.[11]
- Issue-age-rated. Here, the premiums an individual is charged are a function of their age when the policy was first issued. In this case, an individual’s annual premium increases do not reflect the impact that aging has on their expected claim costs, so when a policy is issued the insurer would expect to see a policy’s loss ratio increase steadily over time as the insured ages.
- Community-rated. Here, the premiums an individual is charged are independent of their age; instead, all individuals in a block are charged a common premium. Theoretically, the annual premium increase for the block reflects the impact that the yearly change in the average age of the block is expected to have on the block’s claim costs.
Thus, viewed at the level of the entire block, the insurer may expect at issuance that the block’s expected loss ratio will remain flat over time. However, viewed at the level of an individual policy, the policy’s expected loss ratio can be expected to climb over time, as the block-wide premium increases may not keep pace with the impact of aging on any particular individual’s expected claim costs.
In essence, for a community-rated block, cross-subsidization occurs within the rating structure, with the presence of younger policyholders helping keep the premiums lower for older policyholders than they would otherwise need to be given the insurer’s targeted level of profitability. This will change over time as the block enters the various stages of growth, maturity and decline.
Claim Liabilities
For Medicare Supplement, unpaid claim liabilities are usually estimated in bulk for a block of business using tools and techniques similar to those used for short-duration medical insurance claim liabilities. That is, claims lag triangles are used to determine completion factors, which are then applied against paid-to-date claims to estimate ultimate incurred claims, and then the paid-to-date claims are subtracted from the estimated ultimate incurred claims to obtain the outstanding unpaid claim liability. This contrasts with many other long-duration health insurance contracts, like IDI or LTC, for which most of the unpaid claim liabilities are calculated on a seriatim (claimant-by-claimant) basis using tabular methods.
Commission Structures
With Medicare Supplement, it is common for commissions to be paid out over the lifetime of the policy, and to be calculated as a percentage of initial premium rather than current premium, so that the annual premium increases do not impact the broker’s compensation. Moreover, it is common for the commission percentage to be higher in the early years of the contract, before settling down to an ultimate rate. A typical commission structure might look like this: Commissions in each of the first six years are equal to 20 percent of the first-year premium; then, starting in the seventh year, each year’s commissions are equal to 7 percent of the first-year premium.
This structure, combined with the fact that premiums are expected to increase every year, results in a situation where the commission-to-premium ratio for a policy is expected to keep declining indefinitely, even after the point at which the annual amount of commissions has reached a steady state in dollar terms.
Current Medicare Supplement GAAP Practice
This section represents the author’s perception of typical current industry practice for Medicare Supplement under U.S. GAAP. In a nutshell, it is as follows:
- Issue-age-rated Medicare Supplement policies generate GAAP benefit reserves;
- Attained-age-rated and community-rated Medicare Supplement policies do not;
- All types of Medicare Supplement policies generate unpaid claim liabilities, which are estimated separately from the GAAP benefit reserves (if any);
- All types of Medicare Supplement policies generate DAC assets, reflecting the fact that their commission structure is such that the ratio of commissions to current premiums declines over time; and
- Those DAC assets are amortized over the future lifetime of the policy in proportion to premiums.
In particular, under current U.S. GAAP, many Medicare Supplement policies—namely, those that are attained-age-rated or community-rated—are examples of long-duration contracts that have a DAC asset but zero GAAP benefit reserves.
For health actuaries practicing in this space, it is nearly an article of faith that issue-age-rated policies require GAAP benefit reserves, while attained-age-rated and community-rated policies do not. In thinking through the potential implications of LDTI on Medicare Supplement, it is useful to step back and ask: Why is this the case under current GAAP?
With issue-age-rated business, at policy issuance the insurer expects to see an upward progression of loss ratios for each policy over time, because by design future premium increases will not capture the impact that policyholder aging has on claim costs. As such, at issuance the insurer expects that over time there will be a disconnect between the present value of future benefits and the present value of future net premiums, and therefore the insurer locks in a set of assumptions that create positive GAAP benefit reserves.
With attained-age-rated business, on the other hand, at policy issuance the insurer can argue that it does not expect to see an upward progression of loss ratios for each policy over time, because it intends to manage future premium increases in a manner that will keep the loss ratio steady. Thus, the insurer locks in an expectation that there will never be a disconnect between the present value of future benefits and the present value of future net premiums; in effect, the insurer locks in a benefit reserve of zero at all future points in time.
Moreover, in light of the lock-in principle, as future experience evolves after policy issuance the insurer does not need to revisit the matter. It is true that under current GAAP, the insurer must periodically perform loss recognition testing for its attained-age-rated Medicare Supplement business. This is done by performing a gross premium valuation and comparing the resulting gross premium valuation (GPV) reserve against the net GAAP liability, which here is negative, representing the DAC asset. In practice, the GPV reserve is usually more negative than the negative net GAAP liability, so there is no loss recognition event. If there were such an event, the insurer would presumably start by writing down DAC rather than by leaving the DAC alone and setting up benefit reserves where none previously existed.[12] As such, the author has never seen a situation where a loss recognition event led to GAAP benefit reserves being established for attained-age-rated Medicare Supplement policies post-issuance.
Turning now to community-rated business, the argument becomes more subtle. As discussed earlier, for any particular policy the expectation at issuance may be that the loss ratio will increase over time, as the block-wide premium increases will not keep pace with that policy’s age-related increases in claim costs. However, when we look at things at the block level, the same argument crafted above for attained-age-rated business applies: The insurer’s locked-in expectation at issuance is that the block will be managed in such a way that future loss ratios, measured at the block level rather than the policy level, are kept steady.
There are three reasons why an insurer might argue that looking at things at the block level rather than the policy level is appropriate for community-rated business, thus leading to the conclusion that no GAAP benefit reserves are required today. First and foremost, it aligns the accounting with the economics of how the business is managed. Second, it keeps GAAP consistent with statutory accounting, which explicitly endorses the use of the block as the unit of account for contract reserves associated with community-rated health products, resulting in no statutory-basis contract reserves for these products.[13] And finally, there does not appear to be anything in historical GAAP accounting literature that explicitly prohibits the practice.
Looking Forward
Having laid the groundwork in Part I of this article, in Part II we focus on the application of the LDTI paradigm to GAAP accounting for Medicare Supplement.
Statements of fact and opinions expressed herein are those of the individual author and are not necessarily those of the Society of Actuaries, the editors, or the author’s employer. The author expresses his thanks to Rob Frasca, Bob Hanes and Ken Clark for their comments on drafts of this article.
Rowen B. Bell, FSA, MAAA, is a managing director in the Insurance and Actuarial Advisory Services practice of Ernst & Young. He can be reached at rowen.bell@ey.com.