Life insurance as a retirement income tool

Russell DeLibero, Ph.D., CFPa, Wade D. Pfau, Ph.D., CFAb,*
aThe American College of Financial Services, 270 S Bryn Mawr Avenue, Bryn Mawr, PA 19010, USA
bThe American College of Financial Services, 270 S Bryn Mawr Avenue, Bryn Mawr, PA 19010, USA

Abstract

Given its tax-preferential treatment, careful study is warranted to determine whether life insurance
can play an important role in an overall retirement portfolio. This study develops hypothetical
scenarios for different types of individuals with varying ages and distribution periods, while using a
historical outlook to determine the proper structure of a variable universal life insurance policy. We
compare a variable universal life policy to different investment vehicles (both in qualified and
non-qualified accounts) on an after-tax basis to better understand the potential tradeoff for tax-deferral
and insurance fees within life insurance. © 2017 Academy of Financial Services. All rights reserved.

JEL classification: D14; D81; G11; G23

Keywords: Life insurance; Retirement planning; Asset allocation; Retirement spending

1. Introduction

With the largest generation in the United States entering retirement, retirement planning
is growing in prevalence and relevance as baby-boomers worry about how to fund their
retirements. Studies such as Bengen (2004) and Steiner (2014) have focused on how to spend
from savings during retirement. Baseline assumptions in such studies rarely consider the
impact of taxation. However, there remains widespread concern about taxes and their impact
on retirees (McCarthy, 2011; Silver, 2013). Different types of investment vehicles offer
varying tax treatment. Life insurance is a tax preferential vehicle that one can use as a piece
of an overall asset allocation to help satisfy both retirement income needs, and concerns

* Corresponding author. Tel.: �1-610-526-1569; fax: �1-610-526-1569.
E-mail address: wade.pfau@theamericancollege.edu (W.D. Pfau)

Financial Services Review 26 (2017) 221–240

1057-0810/17/$ – see front matter © 2017 Academy of Financial Services. All rights reserved.



regarding taxation. The strategy of using life insurance as a supplemental retirement income
vehicle raises the issue of whether it is appropriate and what opportunity costs may exist.

Prior research (Katt, 2009; Parish, 2014; Resnick & Resnick, 2009) has shown there are
pitfalls to using life insurance in this manner. However, there are problems with past research
approaches, and a goal of this research is to demonstrate a proper structure to analyze the use
of life insurance as a retirement income vehicle. After determining the proper structure and
framework, this research will then compare how life insurance as an asset class compares to
the performance of equities on both a pre- and post-tax basis. This comparison will provide
transparency to the embedded costs and suitability of insurance for individuals. Research in
this area often does not factor in the tax consequences of accumulation and distribution
strategies. One reason is tax situations can vary so dramatically across individuals. This study
will make assumptions regarding tax rates to provide a clearer picture of the outcomes net
of taxes. The analysis of taxation is crucial as tax savings provide one of the largest benefits
to utilizing life insurance as a supplemental income vehicle.

This study develops a framework for properly structuring variable universal life insurance
for supplemental cash flow during retirement. Two of the most important numbers for this
type of study are the rate of return during the accumulation phase and the safe withdrawal
rate during the distribution phase. Establishing a potential proper rate of return and an
appropriate withdrawal rate will help avoid some of the pitfalls from previous studies using
variable universal life insurance and provide an alternative to the growing problem of taxable
income during retirement. After finding sustainable assumptions for accumulation and
distribution rates in a variable universal life policy, the life insurance vehicle cash flow
generation provides comparison data to an investments-only strategy determining the prob-
ability of success that a qualified or non-qualified investment account would be able to
generate for the same after-tax income. The variable universal life policy was chosen to be
the main source of analysis because it most closely represents a comparison to an
investments-only strategy, providing the owner with flexibility in choosing the underlying
investments (from a menu of options) and risk of the policy. The comparison will demon-
strate any advantages of the tax preferential treatment along with whether life insurance is
a viable replacement for, alternative to, or supplement of retirement income.

The key variables in this study include the following: age at initial plan start, gender,
health classification, length of accumulation, length of income, amount of death benefit,
amount of annual premium paid, historical rates of return, costs of insurance, accumulation
rate, and distribution rate. This study will include multiple scenarios for different ages, length
of distribution periods, and health classification, but the base case will use the following
scenario:

Y Age at initial plan start (45),
Y Health classification (preferred non-smoker),
Y Length of accumulation (19 years),
Y Length of distributions (15 years),
Y Amount of annual premium paid ($50k),
Y Male mortality tables.

222 R. DeLibero, W.D. Pfau / Financial Services Review 26 (2017) 221–240



Six key research questions reflect the objectives for this study include: (1) What is a safe
combined hypothetical accumulation rate and distribution rate to assume when life insurance
is intended to serve as a distribution vehicle? (2) What are the disadvantages for using life
insurance as a distribution vehicle? (3) What characteristics determine appropriate candidates
for using life insurance effectively as a retirement cash flow tool? (4) What is the opportunity
cost of using life insurance as an asset class compared with a brokerage account? (5) What
are the tax advantages to using life insurance as an asset class? (6) Is life insurance a
reasonable choice as a supplemental cash flow vehicle for retirement savings?

Many financial advisors within the industry have shied away from using life insurance as
an asset class. Katt (2009), Resnick and Resnick (2009), and Parish (2014) have all studied
life insurance as a distribution vehicle indicating that both academic and industry studies
have shown poor results. Often this was because of being too aggressive with the hypothet-
ical accumulation or distribution rates. With these poor experiences, this potentially useful
and beneficial strategy may become less common or utilized improperly, which can lead to
suboptimal outcomes for clients. This study aims to provide an appropriate framework for
utilizing life insurance as a supplemental income vehicle and then compare it to other
investment options. This comparison will identify the potential positive attributes for the
supplemental income strategy and will demonstrate the proper way to structure the policy to
obtain more beneficial results. The benefit for at least some individuals will be another source
of retirement funds that offers tax advantages. The more sources that an individual can access
for retirement income, the more prepared that individual will be for different future economic
environments and changing tax regimes.

This study will look at the ability to use life insurance as a distribution vehicle and
determine safe hypothetical accumulation and distribution rates to use in conjunction. This
study will also compare using life insurance as a distribution vehicle with other investment
vehicles on both a pre- and post-tax basis. Several issues arise when looking at using life
insurance for this purpose. Life insurance will have different costs for different individuals
based on age, gender, health classification, type of policy, and carrier chosen. There are also
limitations based on the length of the distribution period for the policy. There is less risk and
more income potential for a shorter distribution period. A significant advantage for using life
insurance as a distribution vehicle is the tax preferential treatment. The comparisons made
to other distribution vehicles with different tax treatment will depend on an individual’s tax
rates for ordinary income and investment income. While this study utilizes a specific set of
assumptions, results will vary in practice based on different individual circumstances.

With retirement planning and subsequently tax planning becoming such prevalent topics,
research on efficiencies and different potential approaches can be beneficial. While there are
many different approaches to retirement income planning with a variety of investment
vehicles, tax planning takes these approaches one-step further. Retirement planning and tax
issues can vary widely for everyone, but it is important to look at a large scope of options
to tailor fit these approaches and ensure the vetting and availability of different options. Life
insurance is a vehicle not originally designed with retirement planning in mind, but because
of its tax preferential treatment, it is a potential option. Prior research discusses the accu-
mulation and distribution phases of retirement planning, along with different potential tax
planning strategies. Additional research has discussed life insurance as a retirement planning

223R. DeLibero, W.D. Pfau / Financial Services Review 26 (2017) 221–240



vehicle, but often it has focused on the pitfalls and potential negative ramifications, and
concluded that it is not a viable option. This study will review this prior research and look
to build and expand upon the potential framework and usefulness of utilizing life insurance
as a distribution vehicle for retirement income.

2. Literature review

Retirement planning is more prevalent now than ever before. In past years, U.S. workers
often received a defined benefit pension from their employer and could rely on Social
Security. The shift to defined contribution retirement plans in the workforce has changed the
scope of retirement planning. In addition, the largest generation in the United States is now
entering its retirement years. Retirement planning consists of two phases, the accumulation
phase, which entails gathering assets, and the distribution phase, which refers to using the
accumulated assets to support spending during non-income earning years.

Bengen (2004) studied what a safe withdrawal rate would be during the distribution phase
to avoid outliving assets in an investment portfolio. This research concluded that 4% would
be a safe initial withdrawal rate to use based on the worst-case historical outcome from
varying 30-year periods in the United States. One critical aspect missing from his original
study is a consideration of taxes, which he did discuss in his book in 2006. The safe initial
withdrawal of 4% is a gross number before the payment of any taxes and will be lower than
4% on an after-tax basis. His assumption was that the investments would be in a qualified
account and, thus, benefit from tax deferral. However, depending on ordinary income tax
rates, especially if higher in the future, individuals may not obtain the highest standard of
living possible if more tax efficient options are available. Not all retirement assets will be
inside a qualified account and taxable investments could be subject to taxation on an ongoing
basis, thus reducing the net amount of the 4% safe initial withdrawal rate.

Sumutka, Sumutka, and Coopersmith (2012) use a comprehensive tax model to evaluate
different withdrawal strategies, discussing the three different types of investment accounts:
tax-deferred, taxable, and tax-free. Sumutka et al. state that the withdrawal sequence from the
various accounts will affect overall taxation on the portfolio and find that the optimal
tax-efficient strategy produces withdrawal stability utilizing low withdrawal rates during
early retirement years. Income stability helps to avoid the loss of itemized deductions, the
loss of tax favored long-term capital gains treatment, and the imposition of the AMT. Tax
efficiency comes from a withdrawal sequence of taxable assets, tax-deferred assets, and then
tax-free assets, keeping in mind offsetting tax deductions and tax-bracket management to
avoid higher taxation. Other withdrawal strategies may provide for smaller tax payments, but
this results in lower wealth creation. Optimal withdrawal strategies demonstrate the impor-
tance of tax planning during retirement and the ability to have multiple asset vehicles to
withdraw from during retirement, including tax diversification and asset location.

McCarthy (2011) concludes that taxes will be increasing in the future based on federal
spending and the looming federal deficit. Tax increases would decrease the standard of living
for many retirees. During retirement, this may mean an increased withdrawal rate, which can
shorten the time horizon that assets will last, or force a lower standard of living. McCarthy

224 R. DeLibero, W.D. Pfau / Financial Services Review 26 (2017) 221–240



states that increasing tax burdens creates opportunities for tax planning. The benefits to
minimizing taxes may have a compounding effect and mean the difference between running
out of money or a lower lifestyle, and living as planned. With tax planning being such a
crucial aspect to retirement income planning, the differential taxation of investment vehicles
needs consideration to determine a net distribution to an individual.

Silver (2013) concludes that existing retirement planning concepts may have served many
retirees well, but with changing laws and historically low tax rates, people are looking for
alternative solutions. There are several different types of vehicles for retirement planning,
and they fall into the three broad categories of tax deferred, taxable, and tax exempt. Silver
references a 2010 survey conducted by Lincoln Financial Group that found taxes constitute
31% of retiree expenses and the amount of taxes paid surprised 11% of retirees. Up to 85%
of Social Security benefits can be taxable, and this is a reason planners are looking to
minimize taxable income during retirement. Because of the potential to join the next tax
bracket and the negative impact on Social Security income, taxable assets are no longer as
desired, even if they offer slightly higher rates of returns, if alternative tax advantaged
options are available. This is especially true when people believe taxes will be higher in the
future.

Many people look at permanent life insurance solely for the death benefit it provides, but
life insurance can serve multiple purposes. A major benefit to permanent life insurance, with
implications for retirement planning, is the ability to accumulate equity within the contract.
What truly makes this an advantage is that the equity accumulates on a tax-deferred basis.
Unlike annuities, life insurance distributions are calculated on a first in, first out basis (FIFO).
One can withdraw the funds placed into the contract without paying any taxes. The next step
is the ability to take loans against the policy. The reason loan provisions are important is that
a loan is not a taxable event, even though there is an on-going interest expense. Because of
this, it is possible to extract a significant amount of cash from within a life insurance policy
without paying taxes. This strategy seems simple and beneficial: accumulate equity, with-
draw up to the basis, borrow gains, and never pay taxes because the death benefit will also
be income tax free. This would certainly help solve problems retirees face with income tax
planning during retirement. However, there are many issues to consider with this basic
strategy.

The tax preferential treatment afforded to life insurance is applicable in general to policies
that can accrue cash, and not to any specific type of policy. However, there are several
different ways to accumulate the cash value, depending on the type of policy. Varying types
of permanent life insurance policy have different features and benefits with potentially
different accrual methods. Often, the type of policy that a person selects depends on the
individual’s risk tolerance, desire for control, flexibility, upside potential compared with
guarantees, and general beliefs about financial markets.

An individual with a more aggressive risk tolerance that would rather choose from a menu
of investments has a better option with a variable universal life policy. Variable universal life
policies offer different subaccounts in which one can make investments; thus, allowing an
individual to have equity market exposure. This potential equity exposure allows for greater
upside potential but also carries risk if the investments do not perform as expected. The

225R. DeLibero, W.D. Pfau / Financial Services Review 26 (2017) 221–240



owner of a variable universal life policy can purchase additional guarantees to ensure
payment of the death benefit regardless of market performance.

Katt (2015) discusses the history of the different types of cash value life insurance policies
and the industry’s evolution over time. In the 1990s when interest rates began to decline,
participating whole life (PWL) and universal life (UL) policies became less attractive. This
led to the introduction of variable universal life (VUL) policies. VULs benefited from
investment performance based on stock and bond subaccounts instead of relying on divi-
dends or an interest-crediting rate. However, early on there were abuses by agents in
illustrating too high of hypothetical rates of return. Most likely this resulted from negligence,
but potentially there was an intent to deceive consumers. Regulation capped maximum
illustration assumptions. VULs, like other equity assets, can be dangerous if policy owners
sell the underlying investments at the wrong time.

Traditional whole life policies work well to provide supplemental income on a tax-
advantaged basis. However, non-guaranteed dividends affect cash accumulation, which
minimizes investor control and flexibility. Kriesel (2010) concludes that many investors were
not happy with the conservative gains found in traditional whole life insurance policies. This
led to the creation of UL to take advantage of the higher interest rates of the 1980s. However,
the marketing of UL policies frequently included maximizing the death benefit with minimal
funding. Unfortunately, when interest rates fell, these policies suffered. Variable universal
life policies take advantage of the ability to invest within the policy via mutual funds.
Because of the tax advantages afforded to life insurance, this can create an opportunity to use
variable universal life policies as a Roth IRA alternative without the income restrictions and
contribution limits.

Katt (2013) further discusses how cash value within life insurance policies is different
from a bank account. A client example discusses a situation where an upset client results
from the charging of interest to access the cash value within the policy and feels that
there should not be a penalty to do so. This is one of the large misunderstandings when
it comes to cash value life insurance. Katt indicates that it should not be thought about
as a bank account, but rather as more like an asset that can serve as collateral to be
borrowed against. Katt further explains a phenomenon known as phantom income, which
occurs with the generation of taxable income from the lapsing of a policy that contains
no cash value. Repaying loans can often be the only option for a client that has an
economic benefit since a lapse or surrender of a policy may result in a substantial taxable
event. Another potential issue can be in the form of surrender charges, which most forms
of universal life policies have. While a surrender charge is also known as an early exit
penalty, the calculation on an illustration is the difference between the accumulation
value and the surrender value. Where additional misconceptions and confusion can occur
is if a client reduces the death benefit during the surrender period. Even without a cash
value withdrawal, the surrender charge still takes place for the reduced portion of the
death benefit. There are primarily two different death benefit options for structuring a life
insurance policy and the cash value accumulation plays an important role in both
approaches. Policies most commonly consist of a level death benefit in which the cash
value accumulation does not increase the death benefit, but borrowing against or
withdrawing the cash value decreases the death benefit.

226 R. DeLibero, W.D. Pfau / Financial Services Review 26 (2017) 221–240



The level death benefit approach helps mitigate the increasing cost of insurance structure
by lowering the net at-risk amount to the insurance company. The net at-risk amount to an
insurance company is the difference between the death benefit and cash value within the
policy. Cost of insurance increases as life expectancy decreases. The payment of death
benefit proceeds exists when there is at least $1 of cash value and does not change if there
is a much higher cash value amount. For this reason, in theory it would be beneficial not to
accrue more than necessary. This would require knowing at what age a person will pass
away, which is why typical policies project to age 100 or beyond. However, if there is a
reduction in life expectancy because of major health concerns, a reduction in premiums can
prevent unnecessary cash value accumulation. Managing a life insurance policy around cash
value accumulation and current life expectancy can provide maximum efficiency.

The type of life insurance policy selected dictates the method and capacity for accumu-
lation of cash value. Each contains its own unique risks for accumulating cash. However,
they all also contain another risk that could nullify the previously mentioned tax advantages
of life insurance. If a policy is deemed to be a Modified Endowment Contract (MEC), the
FIFO account method is no longer available and will contain additional penalties for
accessing the cash value before age 59 1⁄2. The determination comes from a “7 Pay test” and
looks to avoid excessive overfunding of a life policy up-front. This needs to be a consider-
ation when cash accumulation is a goal.

An additional risk exists even when there is the proper structure of a life insurance policy
to avoid MEC rules. Commito (2012) concludes that a Tax court and subsequently an
Appeals court both ruled that a case involving the lapse of a life insurance policy that
contained significant policy loans resulted in tax due because of the significant difference in
the loan balance and cost basis. The client argued that discharge of indebtedness should apply
because of the loan outstanding being greater than the total net worth at the time of the
taxable event, rendering the client insolvent. The Tax court ruled differently, stating that this
did not apply because the client was solvent at the time and loan repayment occurred from
the policy values. This ruling now means that in all situations a taxable event will occur if
a policy lapses with a loan balance greater than cost basis.

Further discussion continues about how managing risk is also crucial during retirement. It
is more involved than simply accumulating enough and providing retirement income. An
entire retirement plan can be at risk if additional unexpected medical expenses or an extended
long-term care need occurs. Unexpected market downturns, inflation, and increasing taxes all
pose additional risks. The design of life insurance helps with transferring multiple retirement
risks by creating the availability of additional resources and riders to provide further
flexibility for policy owners. Policy reprojections, also known as in-force illustrations,
provide an understanding of policy longevity incorporating any market returns or interest
crediting received up to the date of the reprojection. In-force illustrations have a similar
layout to the originally provided sales illustration during policy purchase, but will project
policy longevity based on the current interest crediting assumptions, which are usually
different from the original. In-force illustrations are readily available from the insurance
companies and help monitor the status of a policy to provide a more accurate estimate of
income potential during retirement compared with the original projection. Overloan protec-
tion riders help prevent policies from lapsing because of too many distributions, automati-

227R. DeLibero, W.D. Pfau / Financial Services Review 26 (2017) 221–240



cally placing them in a paid-up status and avoiding the unexpected taxable event from
occurring. Other riders are also available for additional risk management against long-term
care, chronic illness, and death benefit acceleration upon terminal illness. With a supple-
mental income objective, variable universal life is the type of policy that many investors
gravitate toward because they can experience market exposure and upside potential within
the tax efficient wrapper of life insurance.

3. Methodology

Life insurance is a tax-advantaged vehicle that can generate supplemental income. Dis-
cussions around the need for a tax advantaged vehicle during retirement such as life
insurance, which offers tax deferred accumulation, along with the potential for tax-free
distributions, warrants research to the viability of such a vehicle. However, discussions have
also identified that when using life insurance with the intended purpose of supplemental
income, agents often used too aggressive of a hypothetical accumulation rate as well as
too aggressive of a distribution rate. This has resulted in large negative consequences for
individuals with the misunderstanding of how the policy works and the taxable conse-
quences should the strategy not work as originally designed. This research will first
determine an appropriate hypothetical rate of return to avoid the overestimation issues
that have occurred in the past. It will then provide a safe withdrawal rate in conjunction
with the determined accumulation rate, once premiums have stopped, to prevent the
withdrawal of too many distributions that causes the policy to lapse with a taxable event.
Utilizing these two determined rates jointly will alleviate the issues that have caused
trouble for this strategy as noted in previous discussions. Overall, this research will show
the proper structure for these vehicles based on historical data and rolling period
simulations. Thus, this will enhance financial literacy on life insurance strategies by
providing clarification on previous misuses and solutions on the correct structure for this
strategy to be successful.

This study will use a quantitative research approach using historical data, combined with
other select variables to develop a new quantitative approach when assessing life insurance
as a retirement asset. The goal is to determine a quantitative value for the accumulation rate
and its associated distribution rate for a set of client circumstances. Historical performance
will determine the accumulation rate, which is the first part of the equation, as it does not
depend on the distribution rate. However, the distribution rate does depend on the accumu-
lation rate when generating the same amount of cash flow (higher accumulation rate, lower
distribution rate, and vice-versa). Cost of insurance is dependent upon several variables,
including gender, age, health classification, and the amount of the death benefit. For this
study, these arbitrarily selected variables are necessary to determine the cost of insurance
component of the policy. The distribution rate will be dependent upon the duration of the
distribution period, cost of insurance, and accumulation rate. Another arbitrarily specified
variable is the length of distributions. Three different life insurance companies will provide
a sample of costs of insurance. For the baseline, the study considers a 45-year old male client.
Starting ages of 35 and 55 will be used for additional scenarios.

228 R. DeLibero, W.D. Pfau / Financial Services Review 26 (2017) 221–240



For this research, we will also assume preferred non-smoker health for the base case, with
an additional scenario utilizing a standard non-smoker health classification. Preferred non-
smoker is often the second-best health rating given by insurance companies. Typically, the
best classification is uncommon and the goal of this research is to be able to have applicable
findings for the widest possible audience.

Meanwhile, length of accumulation determines how long the assets will be able to
accumulate before providing distributions, with longer periods allowing for greater potential
asset growth. For this research, the base case will assume a 19-year accumulation period,
which for the assumed 45-year-old individual, would put retirement age at 65. Additional
scenarios will include a 9-year accumulation period for a 55-year-old individual and a
29-year accumulation period for a 35-year-old individual.

Length of distributions determines how long assets will be able to support cash flow. This
duration directly affects the amount of cash flow generation, along with a correlation of risk
for the policy to lapse. For the base case, this research will assume a 15-year distribution
period, from ages 65 through 79. The intention is for supplemental cash flow during the early
part of retirement. This approach is not intended to provide the core source of retirement
income over the individual’s lifetime. Testing the combination of accumulation and distri-
bution rates also includes monitoring the policy once distributions stop at age 79. The
purpose of testing after distributions have ceased will be to make sure the policy remains
in force and avoids lapsing for an additional 21 years until age 100, to avoid triggering
a taxable event. Additional scenarios will include distribution periods of 20 and 30 years.

The amount of death benefit directly affects the cost of insurance. A higher death benefit
leads to higher costs of insurance. For this research, the goal is to provide supplemental cash
flow and because this is the focus, the study will solve for the minimum death benefit that
does not trigger a MEC. A minimum non-MEC death benefit will minimize the cost of
insurance, while also preserving the tax advantages of the life insurance policy. The
minimum non-MEC death benefit allows for overfunding of the policy, which helps alleviate
concerns about policies becoming underfunded.

The type of death benefit payout directly affects the cost of insurance based on the net
at-risk amount to the insurance carrier. This study will assume an increasing (Option 2) death
benefit payout during the accumulation period, switching to a level (Option 1) death benefit
payout upon the start of the distribution period. An increasing death benefit is the total of a
base death benefit amount plus the cash value, while a level death benefit does not include
the cash value.

The amount of premium paid directly affects with a positive correlation the amount of
insurance, cost of insurance, and potential distributions. This research will assume annual
premiums of $50,000. Additional scenarios of $10,000 and $25,000 annual premiums will be
tested to confirm the linear relationship between premiums and cash flow generation.

Historical rates of return in the equity markets over an extended period will help determine
an accumulation rate suitable to use for variable universal life insurance contracts. The
assumed portfolio will be 100% equity investments, using large-capitalization U.S. stocks.

The cost of insurance has a negative correlation with the total potential for accumulation
and distributions. Age, health classification, and mortality tables determine the cost of
insurance for each individual insurance company. This study will compare the cost of

229R. DeLibero, W.D. Pfau / Financial Services Review 26 (2017) 221–240



insurance and fees from three separate insurance companies with the goal of providing an
accurate representation of the market. It is important to note that the rates for cost of
insurance are subject to increase up to a maximum allowable amount. An increase to the
maximum allowable amount is historically unlikely, but increases and decreases have
occurred. Actuarial pricing and experience dictates cost of insurance, which may cause
higher or lower costs for different ages.

Policy fees have a negative correlation to the total potential for accumulation and
distributions. Policy fees include premium loads, administrative charges, and death benefit
charges. This study will compare the fees and cost of insurance from three separate insurance
companies with the goal of showing an accurate representation of the market.

This study looks to determine a suitable accumulation rate for variable universal life
insurance contracts. This is the hypothetical rate of return to use for illustration purposes
when someone is buying a new life insurance contract. A proper accumulation rate will avoid
setting an unrealistic or too aggressive of a return expectation that results in the policy
becoming underfunded and lapsing, if no corrective action occurs. This study also looks to
determine a suitable decumulation rate for variable universal life insurance contracts. This is
the hypothetical withdrawal rate to use for illustration purposes when someone is buying a
new life insurance contract. A proper decumulation rate will avoid taking too many distri-
butions from the policy and causing it to lapse.

This study will assume that the distributions from the variable universal life policy will
consist of both withdrawals and loans. Withdrawals occur until the full recovery of cost-
basis, after which loans against the policy occur for remaining distributions. All loans will
assume a contractual fixed loan interest rate of 3%. Current pricing dictates the loan rate
used, but has been higher historically and can be higher or lower in the future.

While life insurance in general receives the tax preferential treatment as previously
discussed, this research will focus on utilizing a 100% equity based variable universal life
policy. The 100% equity based policy provides an equity alternative and minimizes fund
expenses and variations within the allocation options between policies. The design of the
policy for the hypothetical participants will be with the intended goal of providing
supplemental retirement cash flow. With this goal, the VUL policy can take on more risk
as it is only a piece of the overall portfolio, rather than being the main source of
retirement income. A key aspect of the policy design is increasing flexibility during
retirement with the intention of being a complement to the rest of the portfolio and not
a portfolio income replacement. This design will lead to providing the minimum death
benefit allowed by IRS standards that avoids becoming a MEC and losing the previously
discussed tax advantages upon distribution. The minimum death benefit design mini-
mizes the insurance costs and looks to meet the stated objective of maximum supple-
mental cash flow from policy distributions.

Policy design is crucial to the success of utilizing life insurance as a supplemental cash
flow vehicle. The minimum non-MEC death benefit minimizes insurance costs and over-
funds the policy as much as allowable by IRS code while keeping the tax preferential
treatment. The overfunding of the policy creates a larger margin of error before poor
performance and underfunding jeopardize the policy. Underfunded life insurance poli-
cies are the primary concern and pitfall experienced when trying to generate supple-

230 R. DeLibero, W.D. Pfau / Financial Services Review 26 (2017) 221–240



mental cash flow. In addition to the minimum non-MEC death benefit, the type of death
benefit payout is also crucial to the design. The structure of the death benefit payout will
also help maximize the income generation and minimize the cost of insurance throughout
the life of the policy.

With the goal of showing an accurate representation of the marketplace, quotes from three
separate insurance companies are used to determine the cost of insurance. Volume of variable
universal life insurance policies sold and financial strength ratings determine the companies
selected. This will provide a good example of carriers likely to implement this strategy as
described in the research. While variable universal life policies typically allow for the
selection of many different subaccounts with varying equity and bond allocations, this study
will assume an all equity allocation maximizing the potential upside accumulation and
distributions. With the all equity allocation assumptions, this research will be using an
aggressive strategy.

For equity returns, this study uses Robert Shiller’s data set of historical returns from
1871 through 2015 for large-capitalization U.S. stocks (Shiller, 2016). Rolling historical
periods created from the data determine the combined accumulation rate and distribution
rates. Statistical analysis will stress test various accumulation and distribution rates to
determine the historical probability of success. This research will define successful
combined accumulation and distribution rates as those that would have had a 95% or
greater probability of success based on historical equity performance. Testing the
combination of accumulation and distribution rates will also include monitoring the
policy after discontinuing distributions beyond age 79. The purpose of testing after
distributions have ceased will be to make sure the policy remains in force an additional
21 years until age 100 and avoids lapsing, which would trigger a taxable event. The
expectation of the statistical analysis are results showing that people implementing life
insurance as a supplemental cash flow vehicle may have been too aggressive with the rate
of return assumptions and the rates should be more conservative.

To conduct this research, the recreation of a variable universal life insurance policy with
different assumed starting points is necessary. First, multiple life insurance carrier illustra-
tions with the same parameters show the different costs and charges within the policies.
The hypothetical variable universal life policy assumes a 100% investment into a large-
capitalization stock index fund. The aggressive allocation will facilitate equity portfolio
comparisons while minimizing variations of fund costs. For the base case in this study, we
assume the life of the policy will last from age 45 through age 100, with accumulation
occurring during the first nineteen years and distributions occurring during the next fifteen
years. Beyond age 79, the policy contains no inflow or outflow of cash, but requires making
sure the policy remains in-force and does not lapse because of insufficient funds potentially
causing a taxable event. This means that policy simulations require 55 years of investment
returns. The simulations for this research assume a starting point for each year from 1871
through 1961. The study creates 91 simulated life insurance policies, each with a different
55-year historical return scenario. Compiling this data into a sample life policy allowed for
the calculation of the maximum cash flow distribution rate to keep the policy in-force
through age 100.

231R. DeLibero, W.D. Pfau / Financial Services Review 26 (2017) 221–240



4. Results

Table 1 demonstrates the “base” illustration using a specific carrier default 9.41% distri-
bution rate with a selected 6.3% accumulation rate. The illustration shows the end of policy
year age of the individual, amount of premium payments each year, the cash account balance,
the administrative fees and charges within the policy, cost of insurance (COI), income, cost

Table 1 Base variable universal life example with a 6.3% accumulation rate and insurance carrier software
default 9.41% distribution rate

Age Premiums Cash value
starting
balance

Fees COI Income Cost
basis

Loan
balance

Cash
value
ending
balance

Death
benefit

46 $50,000 $ - �$6,712 �$1,380 $ - $50,000 $44,548 $1,088,117
47 $50,000 $44,548 �$4,712 �$1,510 $ - $100,000 $93,891 $1,137,460
48 $50,000 $93,891 �$4,712 �$1,653 $ - $150,000 $146,190 $1,189,759
49 $50,000 $146,190 �$4,712 �$931 $ - $200,000 $202,551 $1,246,120
50 $50,000 $202,551 �$4,712 �$1,059 $ - $250,000 $262,327 $1,305,896
51 $50,000 $262,327 �$4,712 �$1,174 $ - $300,000 $325,747 $1,369,316
52 $50,000 $325,747 �$4,712 �$1,282 $ - $350,000 $393,048 $1,436,617
53 $50,000 $393,048 �$4,712 �$1,400 $ - $400,000 $464,463 $1,508,032
54 $50,000 $464,463 �$4,712 �$1,543 $ - $450,000 $540,225 $1,583,794
55 $50,000 $540,225 �$4,712 �$1,719 $ - $500,000 $620,573 $1,664,142
56 $50,000 $620,573 �$2,120 �$1,863 $ - $550,000 $708,585 $1,752,154
57 $50,000 $708,585 �$2,120 �$2,139 $ - $600,000 $801,849 $1,845,418
58 $50,000 $801,849 �$2,120 �$2,482 $ - $650,000 $900,623 $1,944,192
59 $50,000 $900,623 �$2,120 �$2,841 $ - $700,000 $1,005,239 $2,048,808
60 $50,000 $1,005,239 �$2,120 �$3,217 $ - $750,000 $1,116,046 $2,159,615
61 $50,000 $1,116,046 �$2,120 �$3,773 $ - $800,000 $1,233,242 $2,276,811
62 $50,000 $1,233,242 �$2,120 �$4,131 $ - $850,000 $1,357,442 $2,401,011
63 $50,000 $1,357,442 �$2,120 �$4,516 $ - $900,000 $1,489,056 $2,532,625
64 $50,000 $1,489,056 �$2,120 �$4,478 $ - $950,000 $1,629,003 $2,532,625
65 $ - $1,629,003 �$120 �$4,390 �$153,289 $796,711 $1,573,547 $2,388,120
66 $ - $1,573,547 �$120 �$4,381 �$153,289 $643,422 $1,514,607 $2,234,876
67 $ - $1,514,607 �$120 �$4,211 �$153,289 $490,132 $1,452,134 $2,081,632
68 $ - $1,452,134 �$120 �$4,004 �$153,289 $336,843 $1,385,945 $1,928,388
69 $ - $1,385,945 �$120 �$3,758 �$153,289 $183,554 $1,315,848 $1,775,144
70 $ - $1,315,848 �$120 �$3,473 �$153,289 $30,265 $1,241,638 $1,621,900
71 $ - $1,241,638 �$120 �$3,142 �$153,289 �$123,024 �$ 126,715 $1,159,414 $1,467,035
72 $ - $1,159,414 �$120 �$2,734 �$153,289 �$ 288,405 $1,071,536 $1,307,582
73 $ - $1,071,536 �$120 �$2,233 �$153,289 �$ 454,945 $978,653 $1,143,346
74 $ - $978,653 �$120 �$1,664 �$153,289 �$ 626,481 $880,524 $974,182
75 $ - $880,524 �$120 �$1,269 �$153,289 �$ 803,163 $776,633 $832,473
76 $ - $776,633 �$120 �$1,397 �$153,289 �$ 985,146 $666,060 $723,789
77 $ - $666,060 �$120 �$1,606 �$153,289 �$1,172,588 $548,299 $607,674
78 $ - $548,299 �$120 �$1,842 �$153,289 �$1,365,654 $422,869 $483,615
79 $ - $422,869 �$120 �$2,106 �$153,289 �$1,564,511 $289,255 $351,068
80 $ - $289,255 �$120 �$2,401 $ - �$1,611,446 $304,798 $368,050
— — — — — — — — — —
100 $ - $800,873 �$120 $ - $ - �$2,910,451 $851,200 $851,200

This table utilizes the life insurance carrier default distribution rate of 9.41% with a user selected 6.3%
accumulation rate showing the cash flow each year. Displayed are the end of year age, premiums paid, starting
cash value balance, deductions for policy fees and cost of insurance, income received, cost basis, loan balance
upon recovery of the cost basis, ending cash value balance, and death benefit. The starting base death benefit
amount for the policy is $1,043,569. Not displayed but factored into the calculations are the annual investment
returns. Base case for a 45-year old preferred non-smoker male.

232 R. DeLibero, W.D. Pfau / Financial Services Review 26 (2017) 221–240



basis, loan balance, ending balance for the cash account, and death benefit. The age of the
individual for the base case illustrations will remain between 45 and 100. The beginning and
ending cash account balances are crucial to demonstrate the fluctuation of cash value and
impact that the fees, insurance costs, investment returns, and income all have. If the policy has
a positive cash value, the policy will remain in-force and not lapse, preventing a potential taxable
event. If the insured dies, the proceeds of the death benefit pay off the loan and the beneficiary
receives the net death benefit. If a death benefit payment occurs, regardless of amount, there is
no taxable event. In this example, after distributions and loan balance growth, a death benefit of
$851,200 remains at age 100. With the primary focus of this study being supplemental retirement
cash flow, the death benefit aspects of the policy receive little focus.

With the base example being setup utilizing a carrier default distribution rate of 9.41% and
a selected 6.3% accumulation rate, further analysis and testing provided the impact of
different accumulation and distribution rates. With the creation of and ability to manipulate
a base illustration, the historical performance for each potential starting year determined the
maximum distribution rate while keeping the policy from lapsing.

With a goal of providing a safe assumed rate of return for both the overall accumulation
and distribution periods of the life insurance policy based on historical data, this study will
look at a 95% success rate for the combined accumulation and distribution rates. Table 2
summarizes the success rate for each combination of an accumulation and distribution rate
for the hypothetical variable universal life policy with accumulation rates from a selected
range of 6% through 10% and distribution rates from a selected range of 8% through 11%,

Table 2 Historical probability of success for joint assumptions with a 19-year accumulation period, 15-year
distribution period, and non-lapse for an additional 21 years

Distribution rate

8 8.2 8.4 8.6 8.8 9 9.2 9.4 9.6 9.8 10 10.2 10.4 10.6 10.8 11

Accumulate
rate

6 100 100 100 100 100 100 100 100 98.9 95.6 95.6 94.5 94.5 93.4 93.4 92.3
6.2 100 100 100 100 100 100 100 98.9 95.6 95.6 94.5 94.5 93.4 93.4 92.3 90.1
6.4 100 100 100 100 100 100 98.9 95.6 95.6 94.5 94.5 93.4 93.4 91.2 90.1 85.7
6.6 100 100 100 100 100 98.9 95.6 95.6 94.5 94.5 93.4 93.4 91.2 87.9 84.6 79.1
6.8 100 100 100 100 98.9 95.6 95.6 94.5 94.5 93.4 93.4 91.2 87.9 83.5 78 65.9
7 100 100 100 98.9 95.6 95.6 94.5 94.5 93.4 93.4 91.2 87.9 83.5 76.9 63.7 53.8
7.2 100 100 98.9 95.6 95.6 94.5 94.5 93.4 92.3 91.2 87.9 82.4 74.7 61.5 52.7 50.5
7.4 100 100 95.6 95.6 94.5 94.5 93.4 92.3 91.2 87.9 82.4 72.5 60.4 52.7 50.5 48.4
7.6 100 95.6 95.6 94.5 94.5 93.4 92.3 90.1 86.8 80.2 67 59.3 50.5 49.5 48.4 46.2
7.8 96.7 95.6 94.5 94.5 93.4 92.3 90.1 86.8 79.1 65.9 57.1 50.5 49.5 48.4 46.2 44
8 95.6 94.5 94.5 93.4 92.3 90.1 85.7 79.1 65.9 56 50.5 48.4 48.4 46.2 41.8 40.7
8.2 94.5 94.5 93.4 92.3 90.1 85.7 79.1 65.9 53.8 50.5 48.4 47.3 45.1 41.8 39.6 37.4
8.4 94.5 93.4 92.3 90.1 85.7 79.1 64.8 53.8 50.5 48.4 47.3 45.1 41.8 38.5 37.4 34.1
8.6 93.4 92.3 90.1 85.7 79.1 63.7 53.8 50.5 48.4 47.3 45.1 41.8 37.4 36.3 34.1 34.1
8.8 93.4 91.2 85.7 79.1 63.7 53.8 50.5 48.4 47.3 45.1 41.8 37.4 35.2 34.1 34.1 33
9 91.2 86.8 79.1 63.7 53.8 50.5 48.4 46.2 44 40.7 37.4 35.2 34.1 34.1 31.9 30.8
9.2 87.9 79.1 64.8 53.8 50.5 48.4 46.2 44 40.7 37.4 34.1 34.1 33 30.8 30.8 30.8
9.4 79.1 65.9 53.8 50.5 48.4 46.2 44 39.6 37.4 34.1 34.1 33 30.8 30.8 30.8 28.6
9.6 65.9 53.8 50.5 48.4 46.2 42.9 39.6 37.4 34.1 34.1 33 30.8 30.8 30.8 28.6 27.5
9.8 53.8 50.5 48.4 46.2 41.8 39.6 37.4 34.1 34.1 33 30.8 30.8 30.8 27.5 27.5 27.5

10 50.5 48.4 46.2 41.8 39.6 37.4 34.1 34.1 31.9 30.8 30.8 30.8 27.5 27.5 27.5 25.3

Base case for 45-year old preferred non-smoker male.

233R. DeLibero, W.D. Pfau / Financial Services Review 26 (2017) 221–240



all in 0.2% increments. The combined accumulation and distribution rates are cash flow
generation focused and would look different if the intended goal was to provide more
consideration to the death benefit feature.

Across the range of accumulation and distribution rates, multiple combinations allowed
for a 95% or greater historical success. The table shows that within the specified 6% to 10%
accumulation rate range and 8% to 11% distribution rate range, the upper limit for the
accumulation rate is 8%, with anything greater resulting in less than 95% success for the
lowest distribution rate shown of 8%. An 8% accumulation rate combined with an 8%
distribution rate results in a 95.6% probability of success. In addition, the table shows that
within the specified ranges the upper limit for a distribution rate is 10%, with anything
greater having a less than 95% success rate with the lowest accumulation rate shown of 6%.
A 10% distribution rate results from the accumulation rate being 6%. It is also important to
note that a combination of a 7.6% accumulation rate with an 8% distribution rate had a 100%
historical probability of success, which represents the highest accumulation rate to do so.
With a 6% accumulation rate, a 9.4% distribution rate also results in a 100% historical
probability of success.

Upon further testing of the combined accumulation and distribution rates providing at
least a 95% probability of success, the combination that generated the highest supplemental
cash flow from the hypothetical variable universal life policy was a 7% accumulation rate
with a 9% distribution rate. This combination supported distributions of $158,453. The VUL
rates have been determined with an approach like Bengen’s (2004) SAFEMAX 4% rule
approach, utilizing rolling historical equity returns to determine the maximum feasible
numbers for a given historical probability of success.

The base case scenario for this research utilizes $50,000 annual premiums, however,
additional premium scenarios were tested for $10,000 and $25,000 (20% and 50% of the base
case premiums, respectively). These scenarios resulted in 19.2% and 49.5% of the base case
cash flow, respectively. The majority of the life insurance policy expenses are linear with
only the monthly policy administrative expense being a fixed cost. This results in a fairly
linear relationship between premiums and cash flow generation.

In addition to the base scenario of this research study, involving a 45-year old male
in preferred non-smoker health supplementing cash flow for 15 years, further analyses
provide scenarios with the assumption of a standard non-smoker health classification as
well as for varying accumulation and distribution periods. Table 3 summarizes the
success rate for each combination of an accumulation and distribution rate for the
hypothetical variable universal life policy assuming a standard non-smoker health
classification with accumulation rates from a selected range of 5% through 10% and
distribution rates from a selected range of 7% through 10.75%, all in 0.25% increments.
The change in health classification from preferred health to standard health resulted in
a 0.50% decrease, 7% to 6.5%, in the accumulation rate with the same 9% distribution
rate. This combination supports annual distributions of $145,210, a decrease from the
$158,453 annual distributions with the assumed preferred health. For individuals in poor
health, this may be a less viable strategy based on the higher costs of insurance. A
standard health rating results in a �8.4% cash flow reduction within the given param-
eters, with further reductions expected for substandard health.

234 R. DeLibero, W.D. Pfau / Financial Services Review 26 (2017) 221–240



Table 4 summarizes the corresponding accumulation and distribution rates for a 35, 45,
and 55-year-old male in preferred non-smoker health, with 15, 20, and 30-year distribution
periods. It is important to note that retirement stays at age 65, which is the first year a
distribution occurs. Shorter accumulation periods and longer income periods both work to
reduce the combined feasible accumulation and distribution rates.

To compare the life insurance design to a more traditional investment portfolio, examples
will use the same specified parameters, including amount of investment, and length of
accumulation and distribution periods. Utilizing the same historical equity returns, a prob-
ability of success for generating the same after-tax distributions inside both a qualified and
non-qualified account determines whether the tax advantages overcome the life insurance
policy expenses. There are several different types of investment accounts, because they vary

Table 3 Historical probability of success for joint assumptions with a 19-year accumulation period, 15-year
distribution period, and 21 years of additional non-lapse for a 45-year old nonsmoker with a standard health
classification

Distribution rate

7 7.25 7.5 7.75 8 8.25 8.5 8.75 9 9.25 9.5 9.75 10 10.25 10.5 10.75

Accumulate
rate

6 100 100 100 100 100 100 100 100 100 98.9 95.6 94.5 94.5 93.4 93.4 91.2
6.25 100 100 100 100 100 100 100 100 98.9 95.6 94.5 94.5 93.4 93.4 91.2 87.9
6.5 100 100 100 100 100 100 100 98.9 95.6 94.5 94.5 93.4 93.4 91.2 87.9 80.2
6.75 100 100 100 100 100 100 98.9 95.6 94.5 94.5 93.4 93.4 91.2 85.7 79.1 67
7 100 100 100 100 100 98.9 95.6 94.5 94.5 93.4 93.4 91.2 84.6 79.1 63.7 52.7
7.25 100 100 100 100 100 95.6 94.5 94.5 93.4 93.4 91.2 83.5 75.8 63.7 51.6 49.5
7.5 100 100 100 100 95.6 94.5 94.5 93.4 93.4 90.1 83.5 75.8 61.5 51.6 49.5 47.3
7.75 100 100 100 97.8 94.5 94.5 93.4 93.4 90.1 83.5 73.6 58.2 51.6 49.5 47.3 45.1
8 100 100 97.8 95.6 94.5 93.4 93.4 90.1 83.5 71.4 57.1 51.6 49.5 47.3 44 40.7
8.25 100 97.8 95.6 94.5 93.4 93.4 90.1 83.5 71.4 57.1 50.5 49.5 47.3 42.9 38.5 37.4
8.5 98.9 95.6 94.5 93.4 93.4 91.2 83.5 71.4 57.1 50.5 48.4 47.3 42.9 37.4 37.4 34.1
8.75 95.6 94.5 93.4 93.4 91.2 83.5 71.4 57.1 50.5 48.4 47.3 41.8 37.4 37.4 34.1 34.1
9 94.5 94.5 93.4 91.2 83.5 72.5 57.1 50.5 48.4 46.2 41.8 37.4 35.2 34.1 33 30.8
9.25 94.5 93.4 91.2 84.6 75.8 57.1 50.5 48.4 46.2 40.7 37.4 35.2 34.1 33 30.8 30.8
9.5 93.4 91.2 85.7 75.8 58.2 50.5 48.4 46.2 40.7 37.4 35.2 34.1 33 30.8 30.8 28.6
9.75 92.3 87.9 79.1 60.4 50.5 48.4 46.2 40.7 37.4 34.1 34.1 31.9 30.8 30.8 27.5 27.5

10 89 79.1 63.7 51.6 48.4 46.2 40.7 37.4 34.1 34.1 30.8 30.8 30.8 27.5 27.5 26.4

Table 4 Joint assumptions summary for accumulation and distribution rates for different ages and
distribution lengths based on retirement at age 65

Age

35 (29-year
accumulation)

45 (19-year
accumulation)

55 (9-year
accumulation)

Income duration
15 7.50% 8.75% 7.00% 9.00% 6.75% 7.75%
20 7.50% 7.75% 7.00% 7.75% 7.00% 6.75%
30 7.50% 6.75% 7.00% 6.50% 6.75% 6.00%

Accumulation/distribution rates. Each scenario assumes retirement at age 65 which is the first distribution year.
All scenarios continue to monitor and avoid a policy lapse through age 100.

235R. DeLibero, W.D. Pfau / Financial Services Review 26 (2017) 221–240



in the application of taxation. Non-qualified (taxable) accounts can often benefit from
reduced taxation on dividends and long-term capital gains. Qualified accounts benefit from
deferring all taxes until distribution. However, upon distribution all income is subject to
taxation as ordinary income. Tax-free investment accounts, such as a Roth-IRA, receive the
same potential tax-preferential treatment as life insurance with tax-deferred growth and then
tax-free distributions. The Roth IRA is the best investment account from a taxation stand-
point and therefore does not require further analysis. However, it may not be available for
all individuals based on its income and contribution limits.

This study compares life insurance to both a qualified account and a non-qualified account.
The non-qualified account will assume on-going taxation of historical annual dividends, with
the tax-deferral of price appreciation until withdrawals begin. During withdrawals, taxation
occurs upon recovery of the cost basis. While it is not entirely possible to simply withdraw
the cost basis first within a non-qualified account, based on the average cost basis, after
adjusting for dividends, the first eight distributions are tax-free, and then the remaining seven
distributions are taxable. A common approach is to use the annual taxation method for
non-qualified accounts, however, it is very unlikely that a non-qualified account will have
turnover every year such that there would be annual taxation on the price appreciation.
Multiple scenarios of varying taxation rates on the annual dividend and distributions dem-
onstrate the relationship between higher taxation and probability of success. Taxation rates
will range from 15% to 45%, even though the current maximum taxation on dividends and
long-term capital gains is 23.8%. The higher taxation scenarios demonstrate potential
increases in current taxation levels.

Table 5 summarizes the probability of success that a non-qualified account with various
dividend and long-term capital gains taxation rates can generate the same $158,453 after-tax
cash flow as the variable universal life insurance policy for a preferred non-smoker. As
taxation increases, the probability of success decreases. While the variable universal life
insurance policy has a 95.6% probability of success, the non-qualified account has a lower
probability of success with any assumed dividend and long-term capital gains tax rate greater
than 15%. It should be noted that one bias in using historical data ranging from 1871 to 2015,
is that the historical average dividend yield on stocks was higher in the earlier years than in
more recent times. Therefore, this procedure will overstate the gains associated with a
non-qualified account. Table 5 also summarizes the probability of success that a non-

Table 5 Probability of success for non-qualified account to generate matching after-tax income with various
tax rates

Dividend and
LTCG tax rates

Probability of success:
Matching $158,453 spending
for preferred non-smoker

Probability of success:
Matching $145,210 spending
for standard non-smoker

15% 95.6% 100.0%
20% 90.0% 95.6%
23% 84.4% 93.3%
30% 44.4% 77.8%
35% 40.0% 46.7%
45% 36.7% 41.1%

236 R. DeLibero, W.D. Pfau / Financial Services Review 26 (2017) 221–240



qualified account can generate the same $145,210 after-tax cash flow as the variable
universal life insurance policy when assuming standard non-smoker health. With a reduced
stress on the portfolio needing to generate less cash flow, the probability of success increases.
However, any assumed tax rate greater than 20% results in a lower probability of success
than the variable universal life insurance policy.

Depending on the assumed tax rate, there is a lower probability of success for generating
the same after-tax cash flow as the VUL policy. However, that does not mean the VUL policy
is a replacement for taxable investment accounts, as there are significant differences. Most
notable is the cash value within each vehicle after the specified cash flow period has ended.
The life insurance policy will have cash value that may be accessible for additional
distributions, but it is important not to overdraw the policy and cause it to lapse, triggering
a taxable event. Therefore, most of the cash value remaining is not accessible. There is a
death benefit remaining within the VUL policy, however, that does not provide additional
resources to the insured, only the beneficiary, which will have varying levels of importance
depending on the individual. This differs with the non-qualified account because any
additional remaining value has full accessibility should the individual choose to use it, which
provides additional economic benefit.

The qualified account scenarios reflect that all price appreciation and dividends are tax
deferred until distribution, at which time the full distribution receives ordinary income
taxation treatment. Because of the tax deductibility of qualified accounts, the annual invest-
ments are grossed-up based on an assumed preretirement tax bracket. Nine total scenarios
demonstrate varying pre- and post-retirement tax brackets ranging from 35% to 55%.

Table 6 summarizes the probability of success that a qualified account with various pre-
and post-retirement ordinary income taxation rates can generate the same $158,453 after-tax
cash flow as the variable universal life insurance policy assuming preferred non-smoker
health. Within the nine scenarios, only two have a lower probability of success than the VUL
policy, both of which have a higher ordinary income taxation rate during retirement
compared with the preretirement taxation rate. One of the compelling arguments for the life
insurance policy is protection against a rising tax environment. However, if tax rates do not
rise substantially the qualified plan is clearly more beneficial. With six of the nine scenarios

Table 6 Probability of success for qualified account to generate matching after-tax income with various tax
rates

Pre-retirement
tax bracket

Grossed-up
investment

Retirement
tax bracket

Probability of success:
Matching $158,453 spending
for preferred non-smoker

Probability of success:
Matching $145,210
spending for standard non-smoker

35% $76,923 35% 100.0% 100.0%
35% $76,923 45% 95.6% 100.0%
35% $76,923 55% 55.6% 90.0%
45% $90,909 35% 100.0% 100.0%
45% $90,909 45% 100.0% 100.0%
45% $90,909 55% 94.4% 100.0%
55% $111,111 35% 100.0% 100.0%
55% $111,111 45% 100.0% 100.0%
55% $111,111 55% 100.0% 100.0%

237R. DeLibero, W.D. Pfau / Financial Services Review 26 (2017) 221–240



showing a higher probability of success within the qualified plan compared with the VUL,
individuals should maximize their available qualified plans first before utilizing a life
insurance policy as a retirement vehicle. Even in the scenario of a modest increase in taxation
from 35% to 45%, the qualified plan has the same probability of success as the VUL. With
the VUL also incurring more risk than the qualified plan, the VUL should not be viewed as
a replacement, but rather a supplement when additional tax deferral is desired.

Table 6 also summarizes the probability of success that a qualified account with various
pre- and post-retirement ordinary income taxation rates can generate the same $145,210
after-tax cash flow as the variable universal life insurance policy assuming standard non-
smoker health. With a reduced stress on the portfolio needing to generate less cash flow, the
probability of success increases, resulting in only one scenario of the nine that has a lower
probability of success than the VUL policy. The lone scenario with a lower probability of
success results from a 20% increased tax rate during retirement. As previously noted with the
non-qualified account, the qualified account would also provide additional flexibility to
withdraw and deplete any remaining values if an individual chooses to do so. However, any
additional distributions are taxed at ordinary income taxation rates.

5. Conclusions

This study has determined that for the given parameters of the base case, the safe
combined accumulation and distributions rates are 7% and 9%, respectively. These rates
assume a 100% equity allocation towards U.S. large capitalization stocks. Although other
combinations of rates provided a 95% or higher probability of success based on historical
returns, this combination yielded the highest supplemental cash flow with a minimum of 95%
probability of success. These parameters demonstrate success based on historical perfor-
mance, but in practice it is crucial to monitor the policy performance on an on-going basis.
One of the largest benefits of a variable universal life policy is its flexible nature, which
allows for adjustments to be made. This will allow higher or lower distributions to occur
based on actual performance. This study has also determined that if an individual does not
qualify for preferred health, but standard health, the safe accumulation rate based on
historical returns reduces from 7% to 6.5% with the same 9% distribution rate, which results
in a �8.4% decrease in income with the given parameters.

This research shows that while life insurance can generate cash flow, there are additional
risks that could result in a policy lapsing and triggering a taxable event. If an individual is
in a lower tax bracket during retirement, the tax-preferential treatment would be less
advantageous. The higher the taxation assumptions, the more competitive the life insurance
options become on an after-tax basis because of increased efficiency. However, even in the
scenarios where the life insurance policy provides a higher probability of success compared
with the traditional investment accounts, the residual value after the distribution period has
limited use. This is a result of the need to maintain the policy in-force to avoid taxation. The
investment accounts do not impose such limitations and would provide full access to any
residual values available beyond the distribution period. While the insurance policies have a
tax-free death benefit, it is for a beneficiary and not the insured.

238 R. DeLibero, W.D. Pfau / Financial Services Review 26 (2017) 221–240



Life insurance when used as a supplemental cash flow vehicle needs continuous moni-
toring. Using life insurance as an asset class should be a supplement to an existing retirement
income portfolio and, therefore, is not suitable for every individual. Life insurance can add
benefits for individuals in higher income tax brackets, but for individuals in lower income tax
brackets or individuals who do not have other income solutions, life insurance may not be
appropriate. The reason it is not appropriate for individuals without other income solutions
is that life insurance should not be the primary source of retirement income. In addition,
while the VUL policy is flexible in premiums, concerns could grow if the ability to make
premium payments is no longer possible. While corrective action may be available, such as
reducing the death benefit or surrendering the policy, the payment for costs of insurance may
unnecessarily occur.

While this study focused on utilizing life insurance as a retirement vehicle, it is not
appropriate for everyone. Life insurance is illiquid for generating cash flow and is a
long-term time horizon vehicle. Many policies have a fee or penalty associated with
liquidating the contract during the early years. In addition, because of embedded fees, it often
takes 10 years or more before gains appear within the contract. For these reasons, life
insurance should be a supplement to the overall portfolio for individuals with a long-term
time horizon before requiring income. Individuals considering this strategy should also be
healthy to help balance the costs of insurance associated with the death benefit of the policy.
Individuals who have other retirement income vehicles, are in a higher tax bracket, or believe
that they will be in the future, and can monitor the policy on an on-going basis, may want
to consider utilizing life insurance as a supplemental cash flow vehicle.

The purpose of this research is to show that life insurance is usable as a retirement vehicle
when properly structured. The tax preferential treatment provided to life insurance allows a
consumer to have greater flexibility over which dollars to use during retirement. The cash
accumulation within the policy grows on a tax-deferred basis and then upon retirement,
access to the cash value can occur on a tax-free basis. The life insurance policy provides an
additional option, like the current Roth IRA, but without funding and income limitations.

This research focused on a specific set of parameters to provide supplemental cash flow
to a retirement portfolio. The tax advantages of life insurance as a cash flow vehicle increase
for individuals who are in a higher income tax bracket or individuals who are looking to
protect against future increases in ordinary income tax rates. Tax deferral has been shown to
be beneficial, but some individuals may have limited access to qualified investment accounts.
Life insurance removes limitations to contribution limits. Uncertainty around future taxation
or availability for long-term capital gains treatment may provide reason to look at other
vehicles to supplement a portfolio.

References

Bengen, W. (2004). Determining withdrawal rates using historical data. Journal of Financial Planning, 17,
64–74.

Bengen, W. (2006). Conserving Client Portfolios During Retirement. Denver, CO: FPA Press.
Commito, T. (2012). Court of appeals discusses “Discharge of indebtedness” income when a policy lapses.

Journal of Financial Service Professionals, 66, 14–15.

239R. DeLibero, W.D. Pfau / Financial Services Review 26 (2017) 221–240



Katt, P. (2009). Managing life insurance: Loans or withdrawals. Journal of Financial Planning, 22, 38–39.
Katt, P. (2013). A practical understanding of cash value. Journal of Financial Planning, 26, 40–41.
Katt, P. (2015). History of cash value life insurance and implications for existing policies. Journal of Financial

Planning, 28, 33–34.
Kriesel, W. (2010). Creative uses of insurance. CPA Journal, 80, 50–56.
McCarthy, E. (2011). With higher taxes looming, are your clients ready? Journal of Financial Planning, 24,

22–26.
Parish, S. (2014). The use of life insurance in retirement income risk planning. Journal of Financial Service

Professionals, 68, 33–35.
Resnick, J., & Resnick, B. (2009). Understanding the policy details of no lapse guarantee and VUL. Journal of

Practical Estate Planning, 11, 23–30.
Shiller, R. (2016). Online date Robert Shiller (available at http://www.econ.yale.edu/�shiller/data.htm).
Silver, G. (2013). Tax diversification in retirement planning. Journal of Financial Services Professionals, 67,

49–55.
Steiner, K. (2014). A better systematic withdrawal strategy: The actuarial approach. Journal of Personal Finance,

13, 51–56.
Sumutka, A., Sumutka, A., & Coopersmith, L. (2012). Tax-efficient retirement withdrawal planning using a

comprehensive tax model. Journal of Financial Planning, 25, 41–52.

240 R. DeLibero, W.D. Pfau / Financial Services Review 26 (2017) 221–240