This article first appeared in the Ohio Domestic Relations Journal in two parts:
Vol. 36 No. 6 (November/December 2024) and Vol. 37 No. 1 (January/February 2025)
All domestic relations attorneys have present values float across their desk or computer screen at times. They are dangerous creatures that can bite even the most experienced practitioners. The reason they are dangerous is that actuarial statistics, mortality tables and the details of defined benefit pension plans are not the province of most DR attorneys. But there is another group of attorneys who have a handle on all three areas of knowledge: ERISA attorneys. Fortunately, we interact with them daily and we acknowledge the help of ERISA attorney James Myers (jmyers@qdrogroup.com) in this article.
This article is meant to provide DR attorneys with relevant information for two important practice issues. First, they can assess whether the methodology employed in a present value calculation comports with professional actuarial standards. Second, they can impeach opposing experts who use private brand present values in court.
To successfully deal with present values, knowledge of two elements is essential: the interest rates and mortality tables employed. Those are the two discounting factors used in present values. Each future payment is discounted by an interest factor to reflect the time value of money and a mortality factor to reflect the probability that the payee will survive to receive the payment.
On Thursday, June 6, 2024 the Pension Benefit Guaranty Corporation announced its final changes to “the interest, mortality, and expense assumptions used to determine the present value of benefits for a single-employer pension plan…”1 Subpart B of the PBGC’s regulation on Allocation of Assets in Single-Employer Plans has come into widespread use in the domestic relations community because its stated purpose is to match the cost of replacement annuities in the group annuity marketplace.
The PBGC has a strong incentive to match those annuity market prices. Its charge is to step in and assume the obligations of seriously underfunded pension plans and it must accurately measure its liabilities. As the final regulations explain: “This policy ensures that for a plan entering PBGC trusteeship, the plan’s benefit liabilities are measured consistently with annuity market pricing.”2
The reason that so many domestic relations courts, experts and firms in the field have turned to what we will call “the PBGC Method” to measure the present value of pensions rests on two solid pillars. The first is that courts are tasked with finding the Fair Market Value (FMV) of assets such as pensions and real estate. What better measure of FMV is there than the cost of a replacement annuity in the insurance marketplace? The second element in the popularity of the PBGC Method is the credibility gained by having an objective government arbiter track those values rather than “hired guns” entering DR courts as advocates rather than experts and attempting to peddle a private “made to order” brand of present values.
The changes were intended to improve the transparency and accuracy of measuring those plan liabilities. As an editorial aside we should mention that our quick preliminary analysis of 30 individuals indicates that aggregate liabilities for plans under the changes are unlikely to deviate dramatically from the older PBGC regime. In fact, we3 only found a 1.5 percent difference which is a testimonial to the expertise of PBGC’s actuaries who adjusted the old system with an outdated mortality table to come up with very similar aggregate plan liabilities.
There were five bullet points in the final regulations that summed up the changes:
Before we proceed, it is important to note that the choice of interest rate is by far the most significant factor in determining the present value. The choice of an arbitrarily high or low interest rate for advocacy purposes, even when tied to a published index, leaves the report open to attack and the expert witness vulnerable on cross examination.
To fully understand the PBGC changes we need to look back at how the old present values were calculated with a special focus on how neither the interest rates nor the morality tables used by the PBGC specifically matched industry practice. Instead, PBGC actuaries were clear that neither the interest rates nor the mortality tables used in the present values could be viewed in isolation. Instead, the PBGC actuaries took the sample data obtained from periodic surveys by the American Council of Life Insurers for annuity prices for a wide range of ages and both genders.
The actuaries then drew best fit lines through that data and adjusted the interest rates to best reflect annuity prices. However, in such a system some data points (annuity prices for a specific age and gender) may be further away from the line through the data and thus less reflective of industry practice than other data points.
The new changes eliminate that confusion and misleading criticism that the PBGC interest rates and mortality tables were –in isolation – inconsistent with annuity issuers. The changes will dramatically increase transparency and accuracy across all ages but especially those data points farthest away from the best fit line.
Previously, the PBGC would offer two or three interest rates to be used for present values across all ages. As we explained, that was an indirect approach to match the marketplace. Our past research compared actual annuity prices with PBGC present values and revealed that the values for some ages and genders deviated from industry prices more than others. Or as the PBGC explains “the select and ultimate structure of PBGC’s interest assumptions under the benefits valuation regulation has become increasingly obsolete.”5
Now the PBGC will offer 60 interest rates reflecting the yield curve for different durations ranging from 6 months to 30 years in half year increments. Again, the PBGC points out the reason for the change: “A yield curve approach better reflects the term structure of the fixed income investments that underlie the price of group annuities.”6 It also improves the inherent lag time between the data used to set the PBGC interest assumptions and the valuation date.
How the interest rates are set is now more transparent and “incorporates publicly available bond yield data.”7 Those rates come from the TNC and HQC yield curves but the full and exact methodology in determining the “spreads” has not been released. That means that the core elements of the new yield curve are transparent and predictable but the spreads determined by the PBGC are somewhat opaque.
The interest rates for each time period will be weighted one third to treasury securities and two thirds to high grade corporate bonds with sample yield differences between the two ranging from 12 basis points for short durations and rising to a little over 100 basis points at 30 years. The variance can and will change dramatically in different economic environments.
The reason annuity-writing companies use government and high grade corporate bonds is to protect the company from the vagaries of the stock and other asset markets. If the bond interest payments taken in consistently match the monthly annuity payments that must be sent out, the portfolio can be said to be immunized from loss when a sufficient buffer for administrative expenses is factored in.
Annuity companies that use stocks, junk bonds and other assets to fund annuity payments are playing a precarious game just as annuity companies that promise a share of stock market increases with no possible losses are. Starting in the 1970s actuaries determined through stochastic modeling that overall financial statistics on stock market returns were misleading. While it was true that – on average - things would turn out well, a different picture emerged when astute actuaries ran thousands of simulations through their computers based on random past market performances.
Those thousands of computer simulations indicated the number of times when the company would and would not be able to meet its obligations. Mutual of New York (MONY), for example, found that 17 percent of the time it would fail to meet annuity obligations. Because the odds of failure were nearly one in five, it eschewed the product line at that time.
The new PBGC treasury rates will come from the Treasury Department Nominal Coupon Issues Spot Rates better known as the TNC Yield Curve while the corporate rates will come from the Treasury Department High Quality Market Corporate Bond Yield Curve Spot Rates better known as the HQM Bond Yield Curve. Again, please note the qualifier “high quality” meaning that the corporate bonds are not those being floated by unstable companies and the difference between the treasury securities [bills, notes and bonds] and the corporate bonds do not typically widen significantly until the maturities lengthen.
The blended interest rates will be more market sensitive because they “are based on yields as of the end of the month.”8 However, the final interest rates are not yet set in stone. The blended market yield curve “will be adjusted so that the resulting present values align with group annuity prices.”9
The adjustments or “spreads” will be published quarterly “based on survey data on pricing of private-sector group annuities.”10 This means that the PBGC will, as it has done in the past, first establish a yield curve representing industry pricing in conjunction with the prescribed mortality table. In the past the PBGC often had to make robust changes to the interest rates (typically lowering them) because the mortality tables being used did not reflect current mortality rates.
Clearly, this is a dramatically improved present value system but some additional smoothing takes place. The PBGC determines “the differences between this curve [the blended interest curve adjusted to match market present values] and the blended market yield curve as of the survey date.”11 [Emphasis added] The PBGC then averages those differences “with the differences calculated from prior survey dates to determine the spreads that are used to adjust the applicable blended market yield curve.”12
The initial “spreads” offered in the final regulations were 32 to 38 basis points which were then added to the blended market yield curve which, because of the inverse relationship of interest rates to present values, lowered the resulting present values. This spread adjustment will come in a quarterly table listing 60 adjustments.
The PBGC will timely post on its website13 the yield curve (known as the 4044 yield curve) every month as the data becomes available.
Now we will examine the equally dramatic changes in the PBGC mortality assumptions which now more accurately track insurance industry practice. Mortality assumptions14 “relate to the probabilities that a participant (or beneficiary) will survive to each expected benefit payment date.”15 Mortality assumptions are one of the two essential discounting factors16 in actuarially determining the present value of every possible pension payment. They are based on tables developed from huge data sets recording the longevity of specific cohorts of people. A typical mortality table allows each future payment to be reduced [“discounted”] by the fraction - always less than one – that is the number of people still alive from the original group divided by the number of people alive as of an earlier measurement date.
“Life expectancy” is a concept that is often misapplied in the present value field. It is simply the point in time at which half of the people from the starting group have died. It is important to understand that the actuarial method does not employ life expectancy to calculate present values. Rather, it requires discounting over the entire mortality table from the current age to age 120. The values for each of those years are then summed, after the interest discount, to give the present value. Actuarial Standard of Practice No. 34 at 3.3.4 b. states “…assuming a 100% probability of death at a single age and zero at all other ages is not an appropriate assumption.”17 Yet this is precisely what occurs with the use of a fixed life expectancy, resulting in values that do not reflect how companies price annuities.
Note that same Standard at 3.3.4 c. lends support to the choice of the PBGC method, stating in part “… the actuary may assume the cost of the purchase of an immediate or deferred annuity contract with appropriate benefit features from an insurance carrier.”18 The market based PBGC method has achieved significant success in Ohio with two early court of appeals cases, Conant v. Conant19 and Reitano v. Reitano20 accepting the annuity replacement methodology of the PBGC. In Conant the court stated the expert [QDRO Group’s David Kelley] “then convincingly demonstrated the present cost of an annuity that would pay her the same benefit. That cost is the present value of the pension . . .”
Mortality calculations are onlyas good as the underlying data. Thequest for better data is always the goal of those in the field. More currentand accurate data make for better mortality tables and the recognition thatmortality rates change (usually improving) over time.
The PBGC has adopted more accurate mortality tables with three categories each for men and for women: healthy, disabled, and receiving Social Security disability payments. This makes for a total of six mortality tables along with assumptions on how mortality changes each year. At this writing, the mortality assumption for non-Social security disabled individuals is the same as for healthy individuals.
Here, our astute readers may question the assumption of “mortality improvements” in the face of COVID and the staggering increase in drug overdoses in recent years.21 It is true that those improvements are based on “historical population data through 2019”22 and are somewhat outdated. Others may even point to the relationship between wealth and longevity23 with an eye on the deteriorating financial status of younger Americans compared to their older cohorts emphasizing that “mortality improvements” may not come to pass and those improvements are assumptions on the part of actuaries.
Although the prescribed MP-2021 scales are explicitly referred to ‘mortality improvement scales’, there are some age/year combinations (both in the past and a few in the future) for which the scales return negative improvement rates - that is to say, for some cohorts (notably, some of those attaining an age between the low 20s and the high 40s between now and 2027), mortality is assumed to increase with time rather than decrease. This trend is assumed to be short-lived, with exclusively positive improvement rates for all ages in 2028 and future years, but it is an interesting nuance to the mortality improvement assumption given these scales were developed prior to COVID.
While these points are valid, the PBGC is cognizant of them and is committed to adopting changes as they are needed. In fact, the PBGC promises “to monitor and consider new mortality trend data, including updated mortality improvement scales issued by the Retirement Plan Experience Committee of the Society of Actuaries and intends to amend its regulation to account for new data when appropriate.”24
Actuarial Note: While this explanation may seem esoteric the authors respect our readers and do not wish them to have misleading or superficial understanding of how present values are determined. This in-depth knowledge may prove useful in some situations, especially when present values are litigated. There are two distinct methods for determining the value of those payments: valuing the benefits on a year-by-year basis and then adjusting that value to reflect that the payments are made monthly and actually discounting the month-by-month payments from the onset. The older method is an actuarial adjustment method of calculating the value of an entire year of payments and then modifying those payments based on whether the money is received at the beginning or end of the month.
Because of the computing power of modern systems, it quickly became more commonplace in the insurance industry to forgo the older approximation and instead program monthly discounting. The authors of this article adopted the use of actual monthly calculations years ago for practical considerations to remove a potentially confusing element when testifying about present values in court.
In court, the authors could present pages of printouts of the month-by-month calculations without having to explain the actuarial convention of multiplying the yearly figure by 5/12ths or 7/12ths depending on the timing of the payment. However, that movement towards monthly computer calculations proved prophetic because that is what the PBGC is now prescribing - calculating monthly interest and mortality discounts.
For healthy lives the PBGC is adopting tables in the “Pri-2012 Private Retirement Plans Mortality Table Report published by the Retirement Plan Experience Committee (RPEC) of the Society of Actuaries (SOA) in 2019 (Pri-2012 Report)”25 26 Because the insurance industry “recognizes and distinguishes between mortality for annuitants (i.e. individuals receiving benefits) and non-annuitants (i.e. terminated vested and active participants)”27, the final rule employs non-annuitant tables “for the periods before the participant is projected to commence receiving benefits, and the annuitant mortality tables are used for later periods.”28
“PBGC agrees with IRS and the Treasury Department [who had previously moved to the Pri-2012 tables29] that the Pri-2012 Report is the best available study of the actual mortality experience of pension plan participants (other than disabled individuals).”30
A further refinement is the adoption of generational mortality tables. The basic premise is that a person born in 2022 will live longer than someone born in 2021. Those of us who employ generational tables had to develop far more sophisticated software to take the date of birth of the plan participant into account to generate a mortality rate different from those born in other years.
The mortality table for those receiving Social Security disability benefits has also been updated. The mortality assumptions will “reflect more recent mortality experience by using the table published in the Social Security Disability Insurance Program Disabled Worker Experience Actuarial Study 125…”31
“Multi-factor” mortality is not involved in the new mortality tables. Multi-factor mortality uses multiple data fields [such as the distinction between blue and white collar workers] “to capture the diversity of pension plan mortality by analyzing the characteristics of the individuals in those pension plans.”32
The expense changes are significant because its “current multi-tiered [one rate for liabilities under $200,000 and another for liabilities above $200,000] expense assumptions are too complicated given expense assumptions’ small share of annuity pricing and the simple structure insurers typically use.”33
Instead, the PBGC has set the expense load assumption at a set dollar amount based on the number of plan participants. For present values in a DR case, a $400 load will apply.
Likely not much in most cases. Our initial research of 30 test cases using randomly generated data and comparing the results under the old and new PBGC systems showed results that were quite similar, especially for those closer to retirement.
However, the present values do start to diverge as we look at lives with a longer time horizon to retirement. This is to be expected - increasing the time horizon leads to more variability.
We are taking a careful and comprehensive look at how the present values of the old and the new PBGC factors contrast, as well as a comparison to actual annuity prices in the marketplace - with one caveat. The annuity prices we obtain will be for individual annuities while the PBGC performs a slightly different task.
More important than the effect in dollar terms is the enhanced credibility of the PBGC method. The changes have strengthened the applicability in the present value arena both by increasing accuracy and removing many former grounds for attack.
The authors hope that this article has given the thoughtful attorney a good grounding in the science of present values and the tools to defend objective and expert present values as well as to challenge less sophisticated or advocacy based reports. We gratefully acknowledge the input of Incline Actuarial Group.