A framework for analyzing defined benefit pension
insurance: The survivor benefit plan for veterans

William W. Jenningsa, Jeffrey C. Merrellb, Thomas C. O’Malleyc,
Brian C. Payned,*

aDepartment of Management, United States Air Force Academy, CO 80840, USA
bLeeds School of Business, University of Colorado, 995 Regent Drive, Boulder, CO 80309, USA

cDepartment of Management, United States Air Force Academy, CO 80840, USA
dCollege of Business, University of Colorado-Colorado Springs, 1420 Austin Bluffs Parkway,

Colorado Springs, CO 80918, USA

Abstract

Millions of defined benefit pensioners must select a pension insurance method. We present a
framework for making this decision within the context of U.S. military veterans’ Survivor Benefit Plan
(SBP). Federal government subsidies generate a positive expected net payout for SBP. While
insurance outcomes are typically skewed, the asymmetry of SBP outcomes is stark. In a common
scenario, five percent of participants receive 60% of benefits. An alternative financial planning
approach incorporates private insurance and investments and often bests the SBP. Actuarially correct
life expectancy, moral hazard, taxes, and individual financial needs all play important roles in
selecting a pension insurance program. © 2018 Academy of Financial Services. All rights reserved.

JEL classification: G23; H55; J38

Keywords: Survivor Benefit Plan; Insurance; Valuation; Pension; Retirement; Simulation

1. Introduction

Tens of millions of people in the United States with defined benefit (DB) pensions will
face a key financial decision when they retire: whether and how to insure their DB pension
for their spouse. These insured plans are common among state employees, certain federal

* Corresponding author. Tel.: �1-719-255-3186; fax: �1-719-255-3494.
E-mail address: bpayne3@uccs.edu (B.C. Payne)

Financial Services Review 27 (2018) 147-172

1057-0810/18/$ – see front matter © 2018 Academy of Financial Services. All rights reserved.



employees, and, central to this analysis, military veterans. Richwine (2013) asserts that about
86% of local and state government employees have DB pensions. There are tens of millions
of first responders and state employees as well as over 1.5 million military members who at
retirement can opt for a life insurance product that pays beneficiaries a life annuity upon the
insured’s death. Additionally, although it covers private pensions, the Employee Retirement
Income Security Act of 1974 (ERISA) requires DB plans to provide a survivor’s benefit that
gives pension payments over both the retiree’s and spouse’s lifetimes. These programs
generally have “opt-out” features, meaning couples are defaulted into selecting joint benefits
but can elect not to pay the costs for the prospective survivor benefits. Consequently, these
couples and their advisors need a framework to analyze this key decision.1 This article
provides such a framework using the U.S. military’s pension insurance program as a specific
case. One can tailor the approach presented here to any of the tens of millions facing this
decision.

For military service members the military pension insurance program is called the
Survivor Benefit Plan (SBP). For years the conventional wisdom has been that military
retirees should embrace the SBP. Among the reasons:

Y It requires no medical exam,
Y It is government-subsidized insurance (that implies it is more than actuarially fair),
Y It takes care of a surviving spouse and/or children for their lifetime in real terms, and
Y As Davis and Fraser (2012) summarize, it can provide substantial risk-free real returns.

While these points have merit, the analyses to date are admittedly incomplete. We expand
earlier findings with this study and provide a more granular investigation of the SBP
decision, particularly with respect to military retiree and spouse life expectancies, the
skewness embedded in average returns, and market-based insurance and investment alter-
natives. We also approach this narrow decision from a holistic financial planning perspective
and believe it is imperative that retirees deliberately and thoughtfully consider an SBP-like
decision in light of their complete financial planning and life circumstances.

Based on our analysis, for a broad swath of the retiring military population, we reach the
opposite conclusion of prior studies: that is, we contend SBP is not an ideal approach to
pension protection or life insurance. More strongly, our default position would be to
recommend against SBP unless an individual has a compelling reason or need to select it.2

Because the SBP program is subsidized by the federal government, we posit that other DB
pension recipients could reach the same conclusion when considering their pension insurance
options. Finally, because this conclusion is not the default option during the busy time when
one is transitioning to retirement, we advocate analyzing one’s alternatives in advance of
retirement by two to five years.

Before reaching these overarching conclusions, we map the distribution of potential
outcomes SBP participants can expect to experience and summarize this comprehensive
picture with three major findings:

1. The expected returns are more skewed than previously documented, with more than
60% of retirees in a common scenario expected to pay more into SBP than they
receive, thus, losing wealth;

148 W.W. Jennings et al. / Financial Services Review 27 (2018) 147-172



2. From a holistic financial planning perspective, SBP is largely irrelevant for the
most-likely mortality scenarios; and

3. In circumstances where SBP is most valuable, there exist other life insurance mech-
anisms to provide similar, or even better, financial security.

Our article augments current tools available to federal employees making this critical
financial decision. The Department of Defense Office of the Actuary maintains a SBP
calculator on its website that quantifies the expected benefit based on user-customized inputs
(http://actuary.defense.gov/, “Survivor Benefit Plan” tab). The general takeaway from this
calculator is that SBP provides a financially valuable payout to those who subscribe. In
contrast, this study provides more context for these calculations and for this important
decision. We identify viable alternatives to the SBP and provide a novel calculator that
allows individuals with unique circumstances to compare our proposed alternative and the
SBP (this calculator is located at www.financialcheckpoints.com).

2. Background and literature review

As Poterba et al. (2007) show using a simulation methodology, public defined benefit
pension plans are very generous. According to the analysis presented in Jennings and
Reichenstein (2001), the present value equivalent of a retired military officer pension can
readily exceed $1 million. Currently, military members receive a monthly pension amounting
to 2.5% of their final 36 months’ average pay (nominal) for every year of service.3 For
example, after 20 years of service at 2.5% per year, a retiree receives 50% times their high-36
month average pay, which is usually calculated from the retiree’s last 36 months of pay. For
the typical 20-year enlisted career this amounts to $2,100 per month in 2015; it is approx-
imately $4,000 per month for a typical officer.

Given the longevity of the typical military retirement, these monthly cash flows are well
worth protecting. The life cycle hypothesis holds that households seek stability in consump-
tion and, thus, tend to save during their labor income years and dissave in retirement (Ando
and Modigliani, 1963). With the diminishing marginal utility of wealth, losses are dispro-
portionately harmful. In the event the military retiree dies without SBP protections, the
spouse is not entitled to the retiree’s retirement income. Annuitization stops as does the
associated welfare gains found with annuities in standard life cycle models. This loss
potential is problematic; people who are forward-looking, utility-maximizing, and under-
stand longevity risk should, therefore, choose to protect this income stream to smooth
consumption through time.

To this end, the military’s SBP helps; it protects a portion of the DB pension cash flows
that military members receive after 20� years of service. To protect this pension for the
military member’s family, the SBP permits military retirees to pay up to 6.5% of their
monthly pension pretax to guarantee their beneficiary up to 55% of the monthly retirement
benefit should the military retiree die.4 While the historical opt-in rates for SBP are not
available to the authors, the Department of Defense (DoD) Actuary plans for a majority of
retirees to enroll going forward, using a 50–60% election rate. Of these individuals, most

149W.W. Jennings et al. / Financial Services Review 27 (2018) 147-172



retirees (80%) who elect coverage choose the maximum amount (“Report of the Military
Compensation and Retirement Modernization Commission,” Final Report, January 2015;
Military Compensation and Retirement Modernization Commission 2015). Under the typical
enlisted retiree scenario, the retiree pays $137 pretax monthly for their surviving spouse to
receive a taxable benefit of $1,155 per month. The typical officer’s figures are $260 and
$2,100, respectively.

Unlike the scenario in Lachance et al. (2003), which evaluates pensioners’ option to buy
back into a DB plan from a defined contribution plan, SBP provides very limited optionality.
Similar to many pension insurance programs, the SBP is an opt-out program where the
member and spouse must affirmatively elect to not pay for these benefits. New retirees and
their spouses can opt out immediately or between the second and third year of retirement.
Opt-out decisions are irrevocable, unless the retiree subsequently remarries or adopts a child.
SBP spousal beneficiary payments continue for the remainder of the surviving spouse’s life,
or until the spouse remarries if he or she remarries before age 55. Premiums and payouts are
adjusted with inflation annually. Premium payments cease after 30 years; however, the
coverage continues for the duration of the member’s life. If the spouse predeceases the
member, both premiums and spousal coverage cease immediately. There are no refunds of
premiums paid. For more extensive details on the SBP program, see Higdon (2009).

The SBP pension protection program is essentially an inflation-adjusted annuitized life
insurance product. As such, the analysis here is relevant to many annuity options presented
to public employees. To be clear, the SBP is not an investment, which makes comparing its
return to investments inappropriate. Unlike Comprix and Muller (2011), for example, who
find evidence private firms adjust their DB pension discount rate to suit their purposes, we
are not concerned with a discount rate (i.e., expected investment returns beyond a risk-free
rate) for U.S. government pension assets. Instead, insurance such as SBP operates under the
principle of indemnity. Its intent is to make the insured whole after experiencing a loss. As
with most insurance products, consumers should expect—and in most cases hope—that they
will not have to realize any benefit from paying insurance premiums. Insurance products
exist to prevent harmful large losses. As one author states, “it makes sense to buy insurance
where the premiums will result in a known small loss” (Kitces, 2014). The question this
study examines is whether SBP provides the best “known small loss” option for military
retirees.

While recent work on pension insurance programs is limited (e.g., Bell and Graham, 1984,
analyze 900 private pension plans’ characteristics more than three decades ago), some
analysis of the SBP program has been conducted. Jennings and Reichenstein (2001) dem-
onstrate SBP is actuarially fair under the then-current rules. Using cost-benefit analysis,
Beatty and Kang (2007) question the SBPs value in limited illustrative scenarios. SBP has
been enhanced since these analyses, making it likely that SBP is advantageous for the typical
participant. Davis and Fraser (2012) investigate two changes to SBP that took effect in 2008.
First, a 30-year maximum payment timeframe was then implemented. Second, an offset
feature that reduced SBP benefits for Social Security income was eliminated. The authors
find that with these changes the average implied rate of return to SBP is a riskless 6.8% for
a 45-year-old couple, suggesting the program returns make it quite appealing. Implied returns
are especially generous for male retirees who have younger female spouses.

150 W.W. Jennings et al. / Financial Services Review 27 (2018) 147-172



This study extends the SBP analysis of Davis and Fraser (2012) in ways they recommend
and reaches more muted conclusions. First, we utilize life expectancy tables for military
members rather than using those for the general population. The difference is nontrivial.
Service members are subject to extensive initial health screenings and then mandatory annual
physical fitness tests required to remain in the service. Military members also undergo
routine preventive health exams that sometimes lead to preretirement medically-based
diagnoses that can lead to separation from service. Accounting for these factors leads to a life
expectancy for military retirees (i.e., those who serve 20-plus years) that is significantly
higher than for the general population. The SBP experiences a smaller adverse selection
problem than other insurance programs. Further, given that most military members are not
married to other military members, the typical retiree spouse has a life expectancy more
closely approximated by the general population. As a result, the probability of a spouse
collecting SBP diminishes substantially. In contrast to prior analyses, using military-relevant
life expectancy tables reduces the financial appeal of SBP.

Other contributions of this study address recommendations proposed in Beatty and Kang
(2007). That is, we provide a rich set of life expectancy distribution data, absolute payoff
data, and expected payoff distribution data for a variety of member-spouse mortality
combinations, to include child-only SBP coverage (the online calculator allows myriad age
and status combinations; however, this article depicts the most statistically common scenario
for illustration purposes). We provide this analysis for both the SBP option as well as a viable
cost-neutral alternative, which includes purchasing privately-issued level-term life insurance
coupled with investing the difference between these private premiums and the SBP premi-
ums. Instead of simulating possible member-spouse mortality scenarios, we provide the joint
life expectancy distribution for individual couples. This allows military members to identify
the tradeoffs associated with SBP decision by forecasting a very wide spectrum of possible
outcomes based on their unique circumstances. Furthermore, we present the entire outcome
space, which accurately depicts the proportion of probable SBP “winners” and “losers” for
the first time. This approach enables one to better view the mortality landscape and the
associated skewness in the payoff distribution.

3. Data and methodology

It is important to recognize that on average SBP-insured military officers live longer lives.
Just as the Social Security Administration (SSA) calculates life expectancy tables for the
general U.S. population, the DoD Actuary generates tables for military retirees.5 Fig. 1 shows
that military officer retirees have markedly higher survival rates at every age relative to
general population men and women. While data for first responders are not available to us,
we posit that their life expectancies approximate those of military members more closely
than those of the general population given similarities in health screening, physical training,
job-related physical requirements, and job-related risk.

On the beneficiary side of the equation, it is important to recognize that according to the
DoD Actuary, military spouses unfortunately do not experience the same expected longevity
as the retired military members. Currently, most military retirees are males (90%�),6 and

151W.W. Jennings et al. / Financial Services Review 27 (2018) 147-172



married military members tend to wed non-military spouses (90%�).7 These unbalanced
figures aid our analysis, because the DoD Actuary does not report retiree and spouse tables
by gender. Because the vast majority of spouses are female, we can compare SSA female
probabilities to military retiree spouses. Fig. 1 shows the probability of survival for military
retiree spouses relative to the general population females. Retiree spouses have slightly
higher survival probabilities than SSA females before age 70, but after age 70 they are
comparable. Because military retirees and their spouses differ from the general population,
this warrants an analysis that focuses on their unique life expectancies. Because prior studies
use SSA life expectancy for both military retirees and spouses, they potentially overstate the
SBP value.

To illustrate our method and showcase our results we use a currently common officer
retirement scenario. Our baseline scenario involves a military male retiree age 44 married to
a non-military female who is also 44 years old. At this point a typical officer would have a
20–22 year career, while the typical enlisted member would have a 22–26 year career. The
officer retiree’s insured amount is $55,000. We further assume the spouse does not remarry
once the service member dies. In fact, if a beneficiary spouse remarries before age 55, then
SBP payouts cease unless the remarriage ends in divorce or the new spouse dies. We do
not incorporate this remarriage probability. The value of SBP would decrease relative to
the baseline scenario shown here if remarriage were considered. As for our dollar value
calculations, because future SBP premiums and payouts adjust with inflation, we do not
use a discount factor to adjust them.8 Finally, given our personal biases, we demonstrate the
typical officer retiree calculations, summarize some others, and allow the reader to explore

Fig. 1. Survival probability of U.S. citizens, U.S. military retirees, and retirees’ spouses. This plot shows the
probability of survival by age for an average U.S. male and female according to the Social Security Adminis-
tration. It also shows the probability of survival for U.S. military officer servicemembers who retire and their
spouses, according to the Department of Defense Actuary Office.

152 W.W. Jennings et al. / Financial Services Review 27 (2018) 147-172



alternative scenarios using an online calculator. Obviously, there are many permutations to
this scenario; we summarize a few at the end of this study.

Coupling these assumptions with the SSA and DoD life expectancy tables allows us to
generate a joint probability density function for expected mortality at any age combination
the military retiree and spouse might experience. Fig. 2 depicts this joint distribution for our
age 44 officer retiree scenario, with the highest (or most likely) mortality combination
occurring when the retiree is 88 and the surviving spouse is 89. As expected, the joint
distribution is left-skewed toward lower ages given the rather abrupt upper-bound on
mortality (i.e., the virtual impossibility of living much beyond 110). For ease of interpreta-
tion, most of the subsequent figures incorporate the same basic format and perspective as
Fig. 2.

4. Results

While the joint life expectancy information is necessary, the payouts associated with these
life expectancy combinations is equally important. We calculate these net SBP payouts in
Fig. 3. While we calculate payouts based on $55,000 of retired officer pay, payout calcula-
tions are the same on a percentage basis for any retiree because of the percentage-based
calculations of SBP premiums and payouts. Fig. 3 shows absolute amounts for the typical
officer retiree. It implements the same x- and y-axes as Fig. 2; however, it quantifies spouse’s

Fig. 2. Joint probability of mortality for U.S. military retirees and spouses. This plot shows the joint probability
density function for mortality age of U.S. military retirees and their spouses when both are currently age 44,
according to the according to the Department of Defense Actuary Office. The military member is male; the spouse
is female in this illustration. The highest expected joint probability occurs when the member is age 88 and the
spouse is age 89.

153W.W. Jennings et al. / Financial Services Review 27 (2018) 147-172



payout minus the member’s premiums for each retiree-spouse mortality combination. For
instance, if our 44-year-old military retiree lives to age 50, then he has paid six years’ worth
of SBP premiums, or $21,450. If the officer’s equivalent-aged spouse then lives to age 70,
she collects 20 years’ worth of SBP payouts, or a total of $605,000.9 Taking the difference
between these payouts and premiums yields a net real payout of approximately $583,550 for
the officer who opts for SBP pension protection in this scenario.

Fig. 3 shows who can expect to win and to lose—in strictly financial terms—from the SBP
program. The extreme SBP “winners” are, sadly, the spouses of SBP-covered military
members who die immediately after retiring (as we discuss later, this analysis does not
account for the fact that retirees carry their active duty SGLI [life insurance] coverage for
120 days into retirement at no cost). We hope that the reader understands we put the word
winners in quotations for obvious reasons, as we would rather nobody experiences this
situation! Clearly, the untimely death of a spouse is a personal and financial catastrophe we
wish everyone could avoid. The extreme winning combinations are those in the circled area
of Fig. 3. The net payout value for officer retirees’ spouses in this plot approaches $2 million
in the extreme case where a retiree dies immediately, the surviving widow stays single or
remarries after 55, and then she lives to well over 100.

The net payoff function for the SBP “losers” has a lower bound at 30 years’ worth of
premiums, or $107,250 in our baseline scenario. The maximum negative net payouts—that
represent the worst “losing” situations—occur when the military member pays 30 years’
worth of SBP premiums and the military retiree’s spouse does not survive the retiree. Using
DoD mortality assumptions, this occurs more often than not. In this scenario where the
retiree and spouse are the same age, the probability that the spouse outlives the military

Fig. 3. Survivor Benefit Plan (SBP) net payout (benefits – costs) for U.S. military retirees and spouses. This plot
shows the SBP net payouts, equal to the insurance payout minus premium costs, for various mortality combi-
nations of a military member and nonmilitary spouse who elect to purchase SBP pension protection.

154 W.W. Jennings et al. / Financial Services Review 27 (2018) 147-172



retiree is around 42%, markedly lower than the almost 60% calculated using SSA mortality
tables (Davis and Fraser, 2012). Yet, this result is not surprising given the survival proba-
bility differences shown in Fig. 1. The triangle over the trough in Fig. 3 highlights the
negative payout scenarios. The cost-benefit breakeven point for a spouse to recoup all 30
years’ of premiums is 3.55 years ($107,250 worth of 30-year premiums divided by $30,250
annual payout for officers), meaning a spouse must outlive the retiree by approximately three
and one-half years to recoup all SBP premiums. More generally, for SBP payouts to offset
premiums paid (i.e., breakeven) at any point in time, the spouse beneficiary must outlive the
insured member by 12% of the total duration that the couple has paid SBP premiums.10 The
upward sloping payout on the right side of the triangle represents the mortality combinations
where the spouse predeceases the retiree. In these situations, the retiree has paid some
premiums—but not the full 30 years’ worth—and ceases paying upon the death of his
spouse.

Having laid the groundwork for understanding joint life expectancy and absolute payoffs,
we can now quantify the expected costs versus benefits of the SBP program for an officer
retiree. This is done by simply multiplying the values in Figs. 2 and 3. Doing so yields
Fig. 4, which plots the expected net payout contour for the couple in our scenario.
Visually, the joint life expectancy probabilities dominate the surface relative to the
payouts. Because there is a high probability of both a military retiree and his spouse
living around 80 years, these probabilities tend to dominate the net gains and losses
shown in Fig. 3. The likelihood of “winning” in the SBP—again, the member dying

Fig. 4. Expected Survivor Benefit Plan (SBP) net payout for U.S. military retirees and spouses. This plot shows
the probability-weighted SBP net payouts for various mortality combinations of a military member and nonmil-
itary spouse who elect to purchase SBP pension protection.

155W.W. Jennings et al. / Financial Services Review 27 (2018) 147-172



tragically young and spouse living for a long time—is so low that the multimillion-dollar
payout scenario scarcely registers.

Recognizing that the volume under the curve in Fig. 4 represents the aggregate expected
value of SBP (i.e., Probability � Net payout, where sum of probabilities equals one), some
noteworthy observations result. For instance, as the SBP program and prior authors have
noted, the program is not actuarially-neutral. In fact, its expected net payout (i.e., benefits
minus premiums) for this scenario is over $61,000. Thus, as many claim, SBP is a subsidized
program for the benefit of military retirees. Also, the knife-edge between winning and losing
in the center of the plot stands out. Probabilistically, it is almost a toss-up between whether
(1) the retiree dies early enough for his spouse to collect enough SBP for a positive net
payout (represented by the peak) and (2) the retiree either lives too long for his spouse to
collect enough SBP or that she predeceases him, represented by the deep depression.

We now ask where the SBP value comes from and whether that value most effectively
addresses the retiree’s needs. It turns out that almost 60% of the SBP value comes in an
expanded definition of who are the SBP winners. Specifically, let us expand the definition of
winners to include those scenarios where the military retiree dies before age 66, and his
spouse lives at least 10 years more. In other words, the retiree dies relatively young, and the
spouse then survives for at least another decade. While this particular scenario occurs less
than five percent of the time, it represents almost 60% of the SBP total expected value. Stated
differently, for a common retiree-spouse scenario, non-winners represent over 95% of the
scenarios but only collect 40% of the expected SBP benefit.

The returns to SBP are exceedingly skewed toward those who die relatively young and
have long-lived surviving spouses who do not remarry. A minority of people get more back
than they spend, and a very select few get exceedingly high returns. There is over a 60%
probability that SBP participants will pay more into the SBP program than they ever receive,
despite its federally-subsidized status. This finding surprises us, given that when asked, “Is
SBP a Good Buy?”, the Military Pay and Benefits website sponsored by the Office of the
Under Secretary of Defense for Personnel and Readiness responds, “. . . the answer for most
retirees is yes!”11

Having defined the winner as a low-odds 1-in-20 scenario, we now expand the definition
and determine who are the “usual winners” of SBP. These are couples in which the military
retiree dies after age 66 (i.e., dies later) and the spouse then lives at least 12% longer than
the retiree has been retired (i.e., to get past the breakeven point discussed earlier). Because
these are more likely outcomes, it is critical to consider these cases. By age 67, a beneficiary
spouse will have Social Security options and often other savings (retirement and other), and
perhaps a pension from the spouse’s own job. As a result, the SBP decision becomes one of
marginal wealth. That is, given the couple’s other endowments (savings) and income (Social
Security and pension), will the marginal SBP protection be necessary in the most probable
situations? We contend the answer to this question is: probably not. Thus, our results directly
counter the argument that one should buy SBP because the expected return to SBP is
positive. Note that we do not contend pension insurance is universally a bad idea. Rather, in
contrast to prior findings, we argue that the default position should be against SBP based
strictly on the financial calculations so long as other alternatives exist. We detail some of
these alternatives next.

156 W.W. Jennings et al. / Financial Services Review 27 (2018) 147-172



5. Alternatives to the SBP

If not SBP or any analogous pension insurance, then what? Given the statistical unlike-
lihood of winning the SBP or even experiencing the previously-advertised positive expected
return, what is a (military) retiree to do to protect his or her family? After all, it would be
imprudent to forego pension protection in the early retirement years, which programs like
SBP provide. And the typical military member remains poised and inevitably wants to act on
the information provided here. Fortunately, there are some alternatives to consider.

5.1. Invest for the long term

The first—and most basic—alternative is simply to invest the SBP premiums for long-
term growth rather than using it for this insurance that will most likely never pay out.
According to Davis and Fraser’s (2012) analysis, if a retiree were to earn a real investment
return of 6.8%, he or she would be equally well-off relative to participating in SBP.
However, considering the level of risk required to earn 6.8% real returns on investments
compared with earning these returns in an essentially riskless government benefit (i.e., SBP),
a fiduciary would be hard-pressed to recommend a pure investment strategy over a virtually
riskless SBP strategy.

Relying on uncertain investment returns to pay off sufficiently when needed presents a
risky strategy. Additionally, it simply ignores the nonfinancial utility enhancement life
insurance can provide. We do not recommend this approach because it does not protect the
surviving spouse in the rare but catastrophic situation where the retiree pensioner dies
prematurely. Even if one assumes investment returns that are exceedingly high by historical
standards, portfolio values would likely be insufficient given the short investment horizon.
Indeed, a pure investment strategy would only be able to provide SBP equivalent protection
for a retiree who dies much later in life.

Truthfully, this option represents a so-called straw man. From a holistic financial planning
perspective, it would be imprudent to not protect one’s DB pension in the event of an
untimely death, especially if one’s dependents rely on the pension for nondiscretionary living
expenses. A risky investments strategy is exactly that—risky.

5.2. Buy term and invest the difference

The more responsible option, therefore, becomes the old financial planning adage: “buy
term and invest the difference,” which we call the term-and-invest strategy. One typically
sees this phrase when comparing the purchase of whole life, or cash value, insurance to the
option of buying level-term life insurance. Notably, we do not model a whole life policy
option in this analysis. The reasons for doing so are two-fold. First, in our military retiree
example, these individuals are already conditioned to paying for a term life insurance policy
called Servicemembers’ Group Life Insurance (SGLI) during their time in service. Transi-
tioning to another term policy is a familiar construct. Secondly, and more importantly, we are
considering an illustrative option that is cost-neutral for the military retiree couple when
compared with SBP. Whereas cash value insurance guarantees coverage for life, doing so

157W.W. Jennings et al. / Financial Services Review 27 (2018) 147-172



involves much higher premiums—on the order of 10 times the cost of term life insurance for
the same face value (we depict these premiums in Fig. 6). The typical level-term insurance
we model has a fixed premium for a fixed amount of insurance benefit for a finite period of
time. Term premiums and the insurance coverage amounts are nominally fixed but decrease
over time in real terms. Once the level-term contract expires, if the individual is insurable,
the insured can typically renew a policy for higher premiums, lower death benefit, or a
combination of the two.12

Because both SBP and a term-and-invest strategy provide the utility of having life
insurance, it is not necessary to quantify this value explicitly. Instead, for this analysis,
the relevant considerations are (1) the cost of the SBP coverage, (2) the insurance needs
of the retiring military member, and (3) the cost of these needs from a private insurance
provider. We once again consider the case of our typical officer, a Lieutenant Colonel
(O-5) retiring after 20 years of service at age 44, who has a 44-year-old spouse. Making
some reasonable assumptions for these considerations allows us to quantify this term-
and-invest alternative.

Using an income replacement strategy, foregoing SBP and using private life insurance
means our male officer retiree would need to replace $30,250 in annual SBP income for his
nonmilitary spouse. If we use the well-documented Bengen (1994) four percent payout rule,
the amount of insurance coverage required to provide a real lifetime annuity of $30,250 is
approximately $756,000. Given recent skepticism of the four percent rule as too aggressive,
one might plausibly decrease it to 3.5% or lower. If we use 3.5%, the insurance value
increases to $864,000. For this analysis we use $800,000 in coverage as a happy medium,
which is conveniently the maximum insurance policy our insurance quote source will write
for many active duty officers. We recognize this payout rule is a hotly debated heuristic and
do not intend to provide a recommendation here. Instead, we leave it to an advisor or planner
to conduct sensitivity analyses based on their own philosophies and/or their clients’ situa-
tions. Note that the four percent payout rule and its kin assume inflation adjustments, just as
the SBP does.13

Well-known private companies servicing military members provide level-term life insur-
ance quotes via their Websites.14 A 44-year-old male nontobacco user in good health can
purchase an $800,000, 30-year level-term policy for approximately $125 per month. As-
suming the insurer has sufficient financial strength, this insurance policy can serve the same
need as the revenue stream provided by SBP at less than half the cost of SBP (i.e.,
$125/month vs. $297/month). It is important to acknowledge that insurance payouts are
generally reliable but have more risk than SBP payouts that the federal government backs.
Therefore, we consider two companies’ products. One company has either the highest or
second-highest of 16 or 21 possible ratings from A.M. Best, Moody’s, and S&P; the other
has been in business for over 100 years but does not purchase ratings.

In the long-term, a key part of this buy term and invest the difference option is to invest
the savings. Rather than consuming the $172 monthly savings if purchasing life insurance
instead of SBP, one can save and invest these funds to offset the loss of insurance in 30 years
when the term life contract expires. We cannot overemphasize the fact that maintaining
budget and investment discipline is required for this option to succeed. While it would be
appealing to use an expected market return on an investment portfolio, doing so is not

158 W.W. Jennings et al. / Financial Services Review 27 (2018) 147-172



consistent with finance theory. Although they analyze the pension problem from the payer
side (vs. our focus on the recipient), Brown and Wilcox (2009) cogently argue for using
returns that are risk-equivalent when analyzing investment performance. To ensure we are
comparing risk-equivalent alternatives, we compound the savings component at three percent
annually and use a 2.5% inflation rate to obtain real dollar values from the buy-term-and-
invest strategy.15

Fig. 5, Panel A, shows the present value to the surviving spouse of this term-and-invest
strategy under the aforementioned assumptions for all previously considered mortality
scenarios. In the early years the value of this strategy comes from the level-term policy.
However, because of its nominal $800,000 value, it decreases in real terms until it expires
at 30 years. Correspondingly, the nominal premium also decreases in real terms over time.
The sharp drop off in benefits occurs at the 30-year point when the policy expires. At this
point we see the value of the investments increasing in an exponential manner. To be clear,
we assume all investment contributions cease upon the retiree’s death. However, in the case
of a spouse’s death we assume the retiree continues saving and paying the life insurance
premium since he still receives a pension and might have other desired beneficiaries, such as
children or charities. Also, we see no difference between the winner scenario (i.e., retiree dies
young; spouse lives for decades) and the scenario where both partners die young.
This present value equivalent payout is independent of the surviving spouse’s life span; it
hinges only on when the retiree dies. As we see, dying immediately after the term policy
expires (age 74�) presents the worst situation. At this point the investment amount is
modest, and the level-term coverage is gone. To the extent one can extend the level-term
policy at a reasonable rate or buy a level-term policy that extends beyond 30 years, these
might represent even more appealing options.

The overall expected net payoff for the term-and-invest strategy is $162,214 in our
baseline scenario, $100,946 greater than the expected net payoff ($61,268) of the SBP
option. A subsequent figure summarizes these expected net payoffs for many retiree-spouse
age combinations.

The next figure depicts the horse race, directly comparing payouts to the SBP and the
term-and-invest strategy. Fig. 5, Panel B, expands the earlier figures by differencing Fig. 5,
Panel A, and Fig. 3. That is, it shows the net payout of term-and-invest minus SBP for various
mortality combinations. Anywhere the plot is positive, the term-and-invest strategy bests the
SBP strategy, and vice versa. Fig. 5, Panel C shows the mortality probability weighted
version of the Panel B.

Some valuable observations emerge from these figures. First, when SBP really matters—
that is, if a retiree dies early and the beneficiaries need immediate funds to pay off a
mortgage, and so forth, then the term-and-invest option has a higher present value than SBP
(see circle in Panel B). Second, there is once again a large payout reduction in the
term-and-invest option when the level-term policy expires. This result induces us to recom-
mend that a prospective retiree consider stretching the term policy as far as possible within
reason. Next, there clearly remain mortality combinations where the individual is “less-
covered” by insurance than by SBP, shown in Panel C by the negative expected payout areas
of the contour. However, these points occur later in life, at around age 75, when one would
expect other income streams—like retirement accounts, pensions, or Social Security—to

159W.W. Jennings et al. / Financial Services Review 27 (2018) 147-172



Fig. 5. Panel A: Payout for the buy-term-and-invest strategy. This plot shows, for a variety of mortality combinations,
the present value of payouts for a strategy of buying a term insurance policy and investing the difference between the
cost of the term insurance policy and SBP. The term and invest strategy assumes an $800,000 30-year policy costing
$125/month, 3% return on investments, and 2.5% inflation. Panel B: Payout difference between the term-and-invest
strategy and Survivor Benefit Plan (SBP). This plot shows, for a variety of mortality combinations, the difference in
payouts between SBP and a strategy of buying a term insurance policy and investing the difference between the cost
of the term insurance policy and SBP. If positive, the term/invest policy has a higher payout than SBP. The term and
invest strategy assumes an $800,000 30-year policy costing $125/month, 3% return on investments, and 2.5% inflation.
Panel C: Expected payout difference between the term-and-invest strategy and SBP. This plot shows, for a variety of
probabilistically determined mortality combinations, the difference in payouts between SBP and a strategy of buying
a term insurance policy and investing the difference between the cost of the term insurance policy and SBP. If positive,
the term/invest policy has a higher expected payout than SBP. The term and invest strategy assumes an $800,000
30-year policy costing $125/month, 3% return on investments, and 2.5% inflation.

160 W.W. Jennings et al. / Financial Services Review 27 (2018) 147-172



provide some financial support. Additionally, by investing, a retiree becomes self-insured
much later in life because of the compounded investment returns (see the ramp as the
member’s age increases). Overall, during the highest-probability mortality combinations, the
term-and-invest strategy is preferred over the SBP DB insurance program.

Critically, unlike the SBP, the value of the term insurance and the investments remains intact
even if the spouse dies. That is, the term policy and investments are available to children,
charities, or any future beneficiaries. Besides covering the retiree’s family during the vital early
years, the term-and-invest option provides almost certain non-negative future wealth, regardless
of the mortality order of the retiree and spouse. Finally, and very importantly, these values
represent the marginal benefits (and costs) of the term-and-invest strategy relative to SBP. Even
in those mortality combinations where the term-and-invest strategy underperforms relative to the
SBP strategy, it does not mean the military retiree’s spouse is without financial resources. On the
contrary, Fig. 5, Panel A shows that in these very scenarios the widow does indeed receive
positive payouts in many of the scenarios. The reason these payouts are negative in Panel B is
because the payouts in the SBP winner scenario are exceedingly high. In summary, relative to
SBP, a term-and-invest strategy has many advantages.

To recap the two options succinctly, for the couple identified here, there is a 60%
probability they will pay into SBP more than they ever receive, despite the program’s
government subsidy and despite the program’s expected positive real returns. Conditional
upon falling into this 60% net loss scenario, the couple can expect a net “loss” of $55,232.
On the other hand, if they happen to see any SBP payout, which has a 40% probability of
occurring, their conditional expected payout is $116,500. While this proposition might sound
reasonable in isolation for an insurance product, it loses appeal relative to the term-and-invest
alternative. In the same 60% of the time that the couple would lose with SBP, their
conditional expected payout under term-and-invest is $85,421; for the 40% of the time that

Fig. 5. (continued).

161W.W. Jennings et al. / Financial Services Review 27 (2018) 147-172



SBP pays off positively, term-and-invest pays off as well, with $76,792. All of these figures
involve comparable risk, which is the appropriate way to analyze financial decisions.
Predictably, the differences become even more biased toward the term-and-invest strategy’s
benefit if one posits real investment returns higher than 0.49%.16

Having quantified the most common retiree and spouse scenario and presented the
case for a term-and-invest strategy, it is important to understand there are cases where
SBP clearly makes financial sense. Using the same methodology presented in the
previous scenario, Fig. 6 summarizes other selected officer retiree-spouse age combina-
tions at retirement. With estimates of term insurance premiums and insured SBP
coverage amounts, it provides the expected values of SBP and the term-and-invest
strategy, along with the probability of receiving more from SBP than paid in premiums.
Given the relative life expectancies, it is evident that military retirees with younger
spouses should strongly consider purchasing the SBP program. The shaded combinations
show the retiree-spouse age combinations where the expected payout to SBP exceeds
that of a possible term-and-invest strategy. If a retiree and spouse differ in age by
approximately six years, they will want to look closely at SBP, as eight (four) years
difference makes SBP (term-and-invest) the dominant strategy.

Fig. 6. Expected value of Survivor Benefit Plan (SBP) for various retiree-spouse age combinations at retirement.
This table shows the expected value based on life expectancies for the SBP program available to military retirees
and their spouses, E(SBP). It also shows the expected value for a term insurance plus investment strategy, E(T&I),
where commercial level term life insurance is purchased and the difference between these premiums and SBP
premiums is invested in risk-free treasuries. It also shows the probability that the couple will receive SBP payouts
that surpass premiums, Pr(SBP � 0). The age combinations represent the retiree and spouse ages at the time of
retirement. The final row shows the amount of annual retirement pay the retiree insures. Shaded cells represent
retiree-spouse combinations where the expected SBP payout is greater than the expected term plus investment
payout, or E(SBP) � E(T&I).

162 W.W. Jennings et al. / Financial Services Review 27 (2018) 147-172



5.3. Mentor those junior to you about the SBP decision

For those with clients who have already chosen SBP, it is important to stress that it is not
an ineffective or malicious program. It does provide income protection for the spouse should
the retiree predecease the spouse. SBP is an annuitization strategy that goes a long way in
addressing the longevity risk of a surviving spouse. For many, SBP is a viable solution to the
annuity puzzle, where too few retirees annuitize retirement wealth as life cycle models would
suggest they should (Yaari, 1965). SBP largely addresses an inadequacy in life insurance that
leaves many widows experiencing a dramatic reduction in standard of living (Auerbach and
Kotlikoff, 1991).

With potentially better options available to military retirees, we want to urge mentorship
of future generations about their choices. The FINRA (FINRA Investor Education Founda-
tion, 2012) study entitled, “Financial Capability in the United States, 2012 Report of Military
Findings,” finds that many active military members fail to plan for retirement. Sharing this
SBP analysis with individuals earlier in the preretirement phase of their careers may be
beneficial. In some cases, doing so will allow them to buy level-term insurance at lower rates
and while they are insurable. As good as they are at strategic, operational, and tactical
military planning, many military members allow retirement planning—including the SBP
decision—to flank them. We posit that young workers with private DB plans could benefit
from similar mentoring.

Finally, conflicting incentives converge at retirement, making an early insurance decision
even more important. Typically, military members are incentivized to report every ailment
at retirement for disability qualification purposes under Veterans’ Administration (VA)
benefits programs. Some of these ailments could clearly affect one’s insurability for any
program besides SBP or the military’s term insurance counterpart, Veterans’ Group Life
Insurance (VGLI), neither of which requires a medical exam. Therefore, making the “to SBP or
not SBP” decision earlier could prove tremendously beneficial beyond the figures shown here.
We recommend that service members specifically—and DB recipients more generally—consider
the pension insurance decision two to five years before retirement.

6. Caveats and additional considerations

We have presented a scenario-based analysis using the most common officer retirement
scenario. This means unambiguously that our recommendations are not one-size fits all.
Instead, they come with the following caveats and additional considerations. The online
calculator provides individual-specific insights that this paper does not depict.

6.1. Competent financial planning

First and foremost, this SBP decision is a tactical one that is part of an individual’s and
family’s broader financial strategy. Having a comprehensive strategy is absolutely critical
and the SBP decision should complement this bigger picture. It would be tragic to see

163W.W. Jennings et al. / Financial Services Review 27 (2018) 147-172



destitute war widows and widowers win the tactical financial skirmish but lose the broader
campaign.

6.2. Taxes

As mentioned earlier, our analysis ignores taxes. The basis for this decision is that SBP
premiums are paid pretax, yet benefits are taxed. Conversely, term life insurance payments
are taxed; the proceeds are not. The tax status of investments varies based on whether they
are in tax-advantaged retirement accounts or not, and if so, which flavor (e.g., traditional or
Roth). For these reasons, we urge readers to perform a situation-specific tax analysis before
making any final decisions.17 See the prior recommendation.

6.3. Personal differences

Clearly our scenario-based analysis analyzes a particular couple—again, we understand it
to be the most common officer retiree scenario inasmuch as one exists. Obviously, individual
retiree and spouse health at retirement varies and the couple’s mortality expectations can
deviate significantly—both positively and negatively—from the actuarial table estimates.
For instance, because of a medical condition, it is possible the retiree cannot get private
insurance. One needs to make this determination before declining SBP.

6.4. Age differences

Marked age differences can certainly swing this analysis in favor of the SBP decision,
particularly if the retiree is a male married to a much younger female spouse. In that case it
is more likely that the male retiree will predecease his spouse, generating much higher
expected payoffs to the spousal beneficiary. Fig. 6, as well as our online calculator can help
here.

6.5. Children

Along similar lines as the previous caveat, retirees with young children can consider
opting for the child-only benefit portion of the SBP. This a la carte option permits the retiree
to buy SBP at a very low rate (e.g., on the order of $10–$20 per month) with children as the
beneficiaries. Eligible children share the full 55% benefit equally until they turn 18, or 22 if
a student, at which time they become ineligible. If both spouse and child benefits are elected,
then children are not eligible for these benefits until the spouse is ineligible through
remarriage or death. In short, when a retiree’s youngest child is very young, the SBP child
option can serve as an inexpensive partial substitute for the private insurance coverage in the
term-and-invest option we present. Notably, untabulated analysis shows the child-only
coverage is not actuarially favorable to participants; however, the premiums can be a small
price to pay for the peace of mind this coverage provides.

164 W.W. Jennings et al. / Financial Services Review 27 (2018) 147-172



6.6. The “talk”

As discussed early on, the SBP is an opt-out program. A married member who wishes to
decline coverage must get their spouse’s approval, whereby the spouse recognizes this
decision to forego the annuity upon the member’s death. Anecdotally, we cannot overstate
the difficulties associated with this decision if a couple has not had prior conversations about
their plans. As part of this talk, a spouse must recognize the potential implications of
foregoing SBP. For instance, if a couple declines SBP and subsequently divorces, the spouse
could lose control over the insurance and investments unless they protect themselves
carefully in the divorce proceedings. It is important the spouse recognize this potential
liability. Narrowly, we recommend conversing with one’s spouse early about the SBP
decision. More broadly, we encourage couples to discuss general financial planning matters
early and often!

6.7. Investment and budgetary discipline

Discussing this article with others, including insurance sales representatives, made us
consider putting this item first on the list. As one astute reviewer highlighted, the term-and-
invest alternative depicted here relies on the invest component of the strategy, particularly
later on when returns are compounded over decades. Behaviorally, it takes discipline to keep
investing the difference versus enhancing one’s quality of life. Unfortunately, for some
individuals it is very easy to stop paying a term policy premium or stop investing the
difference if a new car or house upgrade beckons. If a couple is not committed to
investing the difference, then they might consider SBP more carefully. Despite the low
return (i.e., risk-free rate) assumptions in our scenario, the investments represent fully
75% of the term-and-invest strategy’s value. Thus, a term insurance only strategy
underperforms SBP ($40,700 expected payout vs. $61,300 for SBP), which is not
surprising given the SBP subsidy. The real surprise is that a competitively priced life insurance
policy from a reputable insurer competes quite well with SBP at this time. Furthermore, if
allowed to deviate from risk-equivalent returns, a broad market portfolio of investments will
perform much better over decades than shown in our analysis. Additionally, a beneficiary of a
lump-sum insurance payout requires budgetary discipline to ensure the funds last as if they were
an annuity. Financial advisors can assist with this discipline aspect . . . again, see the first
recommendation.

6.8. Assumptions

We make known our assumptions about our scenario, interest rates, and inflation. How-
ever, we must add this item for completeness. Of course, the outcomes are sensitive to these
assumptions. Caveated sufficiently, we do stress that the essence of our recommendations are
not highly sensitive to changes in these assumptions.

165W.W. Jennings et al. / Financial Services Review 27 (2018) 147-172



6.9. Marital status

The younger a widow(er), the more likely he or she is to remarry. Fig. 7 shows these
probabilities. If an SBP recipient widow(er) remarries before age 55, then the SBP benefit
disappears. Consequently, if we added these probabilities of remarriage into this analysis, the
expected payouts to the winners (i.e., the long-lived spouse whose SBP-insured spouse died
young) would decrease. Adding this caveat would on-balance, enhance the appeal of the
term-and-invest strategy relative to SBP. For divorced retirees, Higdon (2009) outlines
multiple family law cases involving SBP. Another application of this analysis results from
these legal situations: divorced service members might find a level-term life insurance policy
a more financially appealing way to provide for their ex-spouse’s insurable interest in the
service member’s retirement pay.

6.10. Credit quality (both ways)

Directly comparing our term-and-invest strategy with SBP assumes the private insurer has
credit quality equal to the U.S. government. We could avoid this assumption by adjusting the
insurance payouts with a discount rate reflecting this risk. We have chosen not to do so at this
time because (1) the private insurers we quote have top ratings and (2) we do not think using
such a discount rate would necessarily obviate any related criticism.

6.11. Female military member

For brevity, this article focused on the most common member-spouse gender scenario;
clearly others exist. Obtaining by-gender mortality data from the DoD Actuary would
facilitate this relevant analysis. Given the relative longevity of military members versus their
spouses along with the typical male-female age differences in marriage we suspect that our
term-and-invest strategy dominates SBP in many female military member cases.

Fig. 7. Probability of a widowed spouse remarriage. This plot shows the probability of a widowed military spouse
remarrying at a given age conditional upon the age at which they were widowed (i.e., 35-, 38-, or 41-years-old).
If remarrying before age 55, a widowed spouse foregoes future Survivor Benefit Plan (SBP) benefits.

166 W.W. Jennings et al. / Financial Services Review 27 (2018) 147-172



6.12. Death Indemnity Compensation (DIC) offset and Servicemembers Group Life
Insurance (SGLI)

Currently, the government’s Death Indemnity Compensation program compensates ben-
eficiaries of military members who die while on active duty or from a disability compensable
under Veterans Administration laws. This DIC currently offsets spousal SBP payments
dollar-for-dollar such that in the aggregate beneficiaries receive up to the maximum of either
DIC or SBP benefits. We have excluded this DIC offset impact in our analysis. However, if
we included it, SBP becomes even less appealing than we portray here, as it is theoretically
possible to pay SBP premiums but never receive an SBP benefit if the DIC maximum
payment subsumes the SBP benefits. Recent military compensation reform proposals rec-
ommend allowing retirees the option of removing this offset; however, in doing so retirees
would pay higher SBP premiums. The net effect can be addressed when it materializes.
Relatedly, servicemembers should know that their SGLI coverage that they carry on active
duty continues for 120 days after retirement at no cost. Currently, the maximum coverage is
$400,000. We do not model this coverage, but if we did, it would temper the relative benefit
of SBP (or strengthen the term-and-invest option) in those scenarios where the servicemem-
ber dies within four months of retirement.

7. Conclusion

Tens of millions of retirees—state employees, federal employees, and even private DB
pensioners—face decisions about whether and how to insure their defined benefit pension.
The U.S. military has a transparent and well-defined pension protection program called the
SBP. The SBP presents a unique opportunity to present a framework for analyzing this key
financial decision. Prior analysis has simulated the payoffs for various scenarios, ultimately
concluding that the SBPs average real return is quite remarkable, between six and seven
percentage for a common scenario. While we concur that this average return makes the SBP
program very attractive, we update the results using more relevant and appropriate life
expectancy data. We show that a 1-in-20 winner scenario is embedded in any rate of return
calculation, masking the highly skewed distribution of outcomes. We ultimately recommend
a default position against the federally-insured pension protection program for the typical
retiree given currently available market alternatives.

One alternative that is currently more appropriate for the typical military retiree considering
SBP-like programs is to buy privately-available level-term insurance and invest the difference
between the term cost and the SBP premiums. Based on an illustration of a common retiree-
spouse combination and risk-equivalent return assumptions, a strategy of buying term insurance
and investing the savings generally provides more preferred outcomes than SBP.

Our analysis couples the Department of Defense Actuary life expectancy tables with realistic
demographic and economic assumptions to first determine the expected payoffs to the SBP
program. It turns out that over 60% of SBP clients should expect to pay more in premiums than
they ever receive. While we might anticipate this breadth of expected negative returns with
private insurers because of profit and overhead requirements, it is somewhat surprising

167W.W. Jennings et al. / Financial Services Review 27 (2018) 147-172



given the SBP is federally subsidized and endorsed. Further, the SBP insures relatively
healthy individuals—U.S. military members who have served at least 20 years—as
evidenced by their required preventive health exams and routine physical fitness tests.

This study further characterizes the expected SBP winners versus typical recipients,
finding that while the winners (i.e., the insured dies young and the beneficiary lives long)
occur only a small fraction of the time (5%), almost 60% of the SBPs expected value goes
to the surviving spouse in these couples (as noted in the paper, we do not mean literal

Fig. 8. Results from Survivor Benefit Plan (SBP) calculator for one common scenario. This figure shows the
inputs available and output summary from a calculator that permits individuals to assess the SBP versus
term-and-invest pension protection strategies given their unique circumstances. Results are based on military-
retiree specific mortality tables. Calculator is available at www.financialcheckpoints.com.

168 W.W. Jennings et al. / Financial Services Review 27 (2018) 147-172



winners. We genuinely hope nobody “wins” early SBP payments). In contrast, the typical
SBP recipients receive their payouts relatively late in the game for relatively few years, at a
time when they likely have other sources of income besides the SBP annuity.

For these reasons, we explore an alternative to SBP, which is buying term insurance and
investing the difference between the cost of the term policy and the SBP premium. This
term-and-invest strategy is superior to the SBP in many situations. It protects the surviving
spouse from an unexpected early death of the insured retiree. Unlike SBP, the term insurance
policy benefits are not conditional upon the beneficiary spouse’s mortality. In other words,
children, charities, or the retiree’s estate will still benefit if the retiree’s spouse passes away.
The investments also provide a certain benefit to the spouse, children, or couple whether the
retiree predeceases them early or not. In fact, in a common scenario of a 44-year-old male
military officer retiring with great health and under plausible nominal investment return (7%)
and inflation (2.5%) assumptions, the expected payout to a term-and-invest strategy is
$330,000, which is over five times the expected payout from SBP. A summary of various
scenarios is found in Fig. 6, which comes from our calculator that we provide online. Fig.
8 shows a snapshot of this calculator.

In closing, it is universally important to recognize certain things about SBP:

Y The expected returns are more skewed than previously documented,
Y Well over one-half of retirees in common situations are expected to pay more in

premiums than their spouse collects;
Y From a holistic financial planning perspective, the most likely mortality scenarios make

SBP largely irrelevant; and
Y In circumstances where SBP is most valuable, there exist other mechanisms to provide

similar, or even enhanced, financial security.

Yet, because of the caveats we list, which include taxes, individual-unique circumstances,
behavioral considerations, our assumptions, and so forth, the SBP decision is truly situation-
specific. Therefore, we recommend readers use this article’s framework as a baseline for
discussing this important topic with their own clients or planners before making a potentially
life-altering decision. We recommend having this conversation two to five years before
retirement. For our part, we are eating our own cooking: those of us who are eligible have
purchased term life policies and declined SBP.18

Notes

1 The analysis in this article clearly assumes that the couple needs to protect the DB
recipient’s pension out of a need for the income replacement. This would not be the
case if, for instance, the couple had substantial wealth (e.g., recent inheritance or
longstanding trust) or income (e.g., spouse earns high income independently) that
would provide for the family’s needs upon the death of the DB recipient.

2 Although the outcome is still an actuarial loss to the military retiree, the one exception
we would recommend is strong consideration of the a la carte child coverage when
military retirees have young children.

169W.W. Jennings et al. / Financial Services Review 27 (2018) 147-172



3 This option is now changing under the new Blended Retirement System. For analysis
of this new option, see Payne et al, 2018.

4 While lesser amounts of coverage for the beneficiary are possible, it is worth com-
paring the benefits of comparable private insurance that we describe later. Details on
this lesser coverage and benefit appear here: http://militarypay.defense.gov/Benefits/
Survivor-Benefit-Program/Costs-and-Benefits/Spouse-Coverage/.

5 Source: “Valuation of the Military Retirement System,” September 30, 2013 and
revised January 2015, produced by the Department of Defense (DoD) Actuary (De-
partment of Defense Office of the Actuary, 2013). Note that we had to extrapolate
mortality probabilities for retirees beyond age 109, as that is the age the DoD Actuary
ended its tables. We assumed a maximum lifespan of 120 years, with exceedingly low
probabilities of living beyond 109. Our extension was informed by SOA RP-2014. We
obtained Social Security data from the Social Sec. Office of the Chief Actuary 2014.

6 http://actuary.defense.gov/, “Statistical Report on the Military Retirement System-
September 30, 2013,” Table: Military Retirees by Gender and Branch of Service as of
September 30, 2013.

7 http://www.militaryonesource.mil/12038/MOS/Reports/2012_Demographics_Report.
pdf, “2012 Demographics: Profile of the Military Community,” Table 2.57: Number
and Percentage of Active Duty Officers and Enlisted Members in Dual-Military
Marriages by Service Branch.

8 Some (e.g., Jennings and Reichenstein, 2003) might argue we should use the TIPS rate
for the real discount rate; however, staying on the other side of that argument makes
our analysis more conservative (i.e., more favorable to SBP than it would be if we used
a TIPS rate). The online calculator permits the user to make this adjustment.

9 For the officer, monthly premiums equal 6.5% � $55,000/12 � $297.92 � 72
months � $21,450. Annual benefits are 55% � $55,000 � $30,250 � 20 years �
$605,000. The enlisted calculations are identical, except they begin with a $2,100 per
month pension benefit.

10 For benefits to outweigh costs, the ratio of spouse lifespan after retiree death/insured
retiree lifespan after retirement must be greater than the ratio of total premiums �
6.5%/total benefits � 55%, which equals 0.12. Thus, the spouse lifespan after retiree
death must be greater than 0.12 � insured retiree’s lifespan as a military pension
recipient and SBP premium payer.

11 http://militarypay.defense.gov/Benefits/Survivor-Benefit-Program/Overview/
12 Because Pfau (2015) makes a compelling case for considering whole life insurance as

a part of holistic retirement planning, we emphasize our recommendation that couples
seek competent financial advice when considering this option.

13 As an astute reviewer pointed out, if one were not comfortable assuming a 3.5– 4%
payout assumption since Bengen (1994) evaluates a risky investment portfolio,
then doing so would require upward adjustments for insurance premiums and
downward adjustments for the invested difference relative to what we present here.
We do not explore these alternatives in this analysis, but they are easy to compute
using the accompanying online calculator. The authors could also provide them
upon request.

170 W.W. Jennings et al. / Financial Services Review 27 (2018) 147-172



14 We also calibrated these costs with a broad group of general life insurance providers
to ensure their reasonableness. The military-centric coverage is more expensive,
presumably because of the consideration of such factors as high-risk work and the
absence of war exclusion clauses. Evaluating military-specific policies is, therefore,
conservative.

15 We assume 2.5% inflation and 3% return on investments for a 0.49% real return on
investments. Doing so is consistent with the Jennings and Reichenstein (2003) ap-
proach to valuing defined benefit pensions. Note the results are relatively less sensitive
to these return assumptions than they are to the insurance coverage amount, term, and
cost.

16 For instance, assuming a reasonable investment portfolio return of 7% (4.39% real),
the conditional expected payouts for the term-and-invest scenario become $125,665
and $205,335 in the 40% and 60% spaces, respectively, for a total expected payout to
term-and-invest of $331,000.

17 Using the illustrative example in the article, we assume a 25% pre-death income
tax rate (federal and state), 17% post-death income tax rate (federal and state), and
15% capital gains rate and that any investments are in non-tax-advantaged ac-
counts. Under these conditions, the expected value of SBP decreases by $2,620 or
4.3% while the expected value of term-and-invest decreases by $3,465 or 2.1%.
The magnitude of the tax impacts is similar under both options. Obviously, there
are countless permutations based on individual circumstances, leading us back to
the first consideration in this list.

18 Full disclosure: one of us with young children opted for the child-only option.

Acknowledgment

The views in this article are solely the authors’ and do not reflect those of the U.S.
Government, Department of Defense, U.S. Air Force, or U.S. Air Force Academy. It was
previously circulated as “Alternate Options to the Survivor Benefit Plan: Important Consid-
erations for US Veterans.” We gratefully acknowledge comments and assistance with this
project from participants at the 2014 and 2017 Academy of Financial Services Annual
Meetings, seminar participants from the Department of Management at the U.S. Air Force
Academy, Lieutenant Colonel (retired) Stephanie Bruce, and Lieutenant Colonel Joseph
Suhajda.

References

Ando, A., & Modigliani, F. (1963). The “life cycle” hypothesis of saving: Aggregate implications and tests. The
American Economic Review, 53, 55–84.

Auerbach, A. J., & Kotlikoff, L. (1991). Life insurance inadequacy-evidence from a sample of older widows
(Working Paper No. 3765). Retrieved from National Bureau of Economic Research website: https://www.
nber.org/papers/w3765.pdf.

171W.W. Jennings et al. / Financial Services Review 27 (2018) 147-172



Beatty, S. E., & Kang, T. (2007). The Survivor Benefit Plan: A Cost-Benefit Analysis MBA Professional Report,
Thesis and Dissertation Collection, Monterey, California. (available at https://calhoun.nps.edu/handle/10945/
10185)

Bell, D., & Graham, A. (1984). Surviving spouse’s benefits in private pension plans. Monthly Labor Review,
April, 23–31.

Bengen, W. P. (1994). Determining withdrawal rates using historical data, Journal of Financial Planning,
October, 171–180.

Brown, J. R., & Wilcox, D. W. (2009). Discounting state and local pension liabilities. American Economic
Review, 99, 538–542.

Comprix, J., & Muller, K. (2011). Pension plan accounting estimates and the freezing of defined benefit pension
plans. Journal of Accounting and Economics, 51, 115–133.

Davis, K., & Fraser, S. P. (2012). Insuring defined-benefit plan value: An examination of the survivor benefit plan
(SBP) decision. Financial Services Review, 21, 275–290.

Department of Defense Office of the Actuary. (2013). Valuation of the Military Retirement System. September
30, 2013, revised January 2015. (available at https://actuary.defense.gov/Portals/15/Documents/MRF%
20ValRpt%202013.pdf?ver�2015-02-24-102857-617)

FINRA Investor Education Foundation. (2013). Financial Capability in the United States: 2012 Report of
Military Findings. (available at http://www.usfinancialcapability.org/downloads/NFCS_2012_Report_
Military_Findings)

Higdon, J. N. (2009). The Survivor Benefit Plan: Its history, idiosyncracies, coverages, cost, and applications.
Family Law Quarterly, 43, 439–488.

Jennings, W. W., & Reichenstein, W. R. (2001). The value of retirement income streams: The value of military
retirement. Financial Services Review, 10, 19–35.

Jennings, W. W., & Reichenstein, W. R. (2003). Valuing defined-benefit plans. Financial Services Review, 12,
179–199.

Kitces, M. (2014). Insurance rule: Buy the losers. Financial Planning, December, 73–76.
Lachance, M. E., Mitchell, O., & Smetters, K. (2003). Guaranteeing defined contribution pensions: The option

to buy back a defined benefit promise. Journal of Risk and Insurance, 70, 1–16.
Military Compensation and Retirement Modernization Commission. (2015). Report of the Military Compensation

and Retirement Modernization Commission. Final Report. (available at https://www.ngaus.org/sites/default/
files/MCRMC%202015_0.pdf)

Payne, B. C., Jennings, W. W., & O’Malley, T. C. (2018). Implied discount rates for the U. S. Military Blended
Retirement System. Journal of Retirement, 6, 22–32.

Pfau, W. D. (2015). Optimizing retirement income by combining actuarial science and investments. OWRS.
(available at http://www.owrsfirm.com/wp-content/uploads/OA_WP_Opt-Ret_Inc_05–15_web.pdf).

Poterba, J. M., Rauh, J., Venti, S., & Wise, D. (2007). Defined contribution plans, defined benefit plans, and the
accumulation of retirement wealth. Journal of Public Economics, 91, 2062–2086.

Richwine, J. (2013). Nine fallacies used to defend public-sector pensions. Backgrounder, The Heritage Foun-
dation, No. 2765, Washington, DC. (available at https://www.heritage.org/social-security/report/nine-
fallacies-used-defend-public-sector-pensions)

Social Security Office of the Chief Actuary. (2014). Statistical Tables, Period Life Table. (available at https://
www.ssa.gov/oact/STATS/table4c6.html)

Yaari, M. E. (1965). Uncertain lifetime, life insurance, and the theory of the consumer. The Review of Economic
Studies, 32, 137–150.

172 W.W. Jennings et al. / Financial Services Review 27 (2018) 147-172