Framing the Annuity as Bequest Protection: An

Experimental Test

Ying Yana, Russell N. James IIIb,*

aDepartment of Personal Financial Planning, Eastern New Mexico State University, 1500 S Avenue K,

Portales, NM 88130, USA
bDepartment of Personal Financial Planning, Texas Tech University, Box 41210, Lubbock,

TX 79409-1210, USA

Abstract

Bequest motives are commonly cited as a barrier for annuitization. This paper tests how framing

partial annuitization as a protection for an intended bequest against the risk of asset exhaustion due

to unexpected longevity influences the desire to purchase an annuity among a sample of 2,160 partic-

ipants. The results indicate that this framing argument does increase interest in purchasing an annu-

ity. Regression results demonstrate that this framing has a larger positive effect for individuals with

a greater bequest motive. The relationship between annuitization framing and bequest motive dem-

onstrated by this experiment has important practical and theoretical implications. © 2021 Academy

of Financial Services. All rights reserved.

JEL classifications: D1; D14; D15

Keywords: Annuities; Estate planning; Bequest motive; Framing effects

1. Introduction

Longevity risk is a risk that a person lives longer than expected and has insufficient

wealth to support his or her planned consumption expenditures. A life annuity is an insur-

ance instrument that converts wealth into a lifetime income stream. It can insure people

against longevity risk and can serve as a valuable part of retirees’ investment portfolios

(Lockwood, 2012). However, relatively few people annuitize any of their wealth

*Corresponding author: Tel.: +1-806-787-5931; fax: 806-742-5033.

E-mail address: russell.james@ttu.edu

1057-0810/21/$ – see front matter © 2021 Academy of Financial Services. All rights reserved.

Financial Services Review 29 (2021) 277–291



(Peijnenburg et al., 2016). Franco Modigliani (1986) referenced this “annuity puzzle” in his

Nobel Prize acceptance speech and indicated that there are substantial reasons for people to

annuitize at least part of their wealth.

Bequest motives are a common explanation for this annuity puzzle. Annuity benefits will

end at the death of the annuitants; therefore, they are not bequeathable wealth. Economists

and researchers indicate that this non-bequeathable characteristic may potentially reduce or

even eliminate the desire for annuitization (Bernheim, 1991; Bütler & Teppa, 2007; Yaari,

1965). Bequest motives, or proxies for such, have been measured in a variety of ways such
as the presence of children, bequest expectations, or the self-reported importance of leaving

a bequest. Typically, researchers do find a negative association between a bequest motive

and purchasing a life annuity (Bernheim, 1991; Bütler & Teppa, 2007; Yaari, 1965).

This paper examines the effects of framing an annuity as bequest protection on intentions to

purchase an annuity and how these effects are impacted by the bequest motive. The estimations

of regression models show that framing the annuity as an inheritance protection did increase in-

terest in using an annuity. Further, it did so to a much stronger degree for people who first

reported having a relatively higher bequest motive. This was true both when measured as the

change in estimated numerical probability of using an annuity (multiple linear regression

model) and when measuring the binary outcome of whether or not interest in an annuity

increased (probit regression model). The increase in the interest in annuitization grew more

strongly for those with a higher bequest motive, even when controlling for other factors such

as the personal longevity estimations, education, income, age, and gender. A higher bequest

motive leads to a greater impact from bequest protection framing of an annuity.

2. Literature review

2.1. The annuity puzzle

People face significant longevity risk in the United States (Lockwood, 2012) in part due

to generally increasing life expectancy (Knell, 2018). According to data from the World

Bank, the average U.S. life expectancy for males increased from 46.3 to 79 between 1900

and 2018 and the average life expectancy for females increased from 48.3 to 84 between

1900 and 2018 (The World Bank, 2019). About one-fifth of 65-years-olds will live to age 90

and beyond (Lockwood, 2018). Therefore, retirees should be increasingly concerned about

the need to hedge longevity risk.

The economic literature provides theoretical and empirical evidence that life annuities

bring substantial welfare improvements to retirees (Cox & Lin, 2007; Scott, 2008). An im-

mediate annuity can bring a lifetime income stream for retirees and help them maximize

retirement spending (Scott, 2008). Moreover, Scott (2008) suggests partial annuitization

should be an optimal strategy for typical retirees and recommends that retirees convert 10-

15% of wealth to a life annuity in the retirement period.

However, few retirees annuitize any of their wealth. Among people who are eligible for

defined benefit pension plans, a form of an annuity, when the plans offer a lump-sum option,

about 50% to 75% of these benefits are taken as a lump-sum (Banerjee, 2013). This is true

278 Y. Yan and R. N. James / Financial Services Review 29 (2021) 277–291



even if the lifetime benefit is the default option and retirees have to spend great time on com-

plex paperwork to get the lump-sum distribution (Banerjee, 2013; Benartzi, Previtero, &

Thaler, 2011; Mottola & Utkus, 2008). Williams and James (2019) found that only 4% of

retirees were receiving income from a commercial annuity other than traditional pension

plans. Thus, the propensity to annuitize is low in the United States (Peijnenburg et al.,

2016). Unwillingness to annuitize is not a unique behavior for Americans. For example, an

analysis using Japanese data also reports similarly low propensity to voluntarily annuitize

wealth (Purcal & Piggott, 2008). This widespread resistance to annuitization is known as the

“annuitization puzzle” (Modigliani, 1986). In other words, annuities are underutilized.

2.2. Bequest motives

Researchers have used various explanations for this annuitization puzzle. The low

demand could result from the public pension system. Retirees may have sufficiently hedged

the longevity risk with annuities provided by the public pension system, such as Social

Security (Brown, 2001, 2003). Others argue that the annuity market is not actuarially fair

(Brown & Poterba, 2000; Donnelly, 2015; Lockwood, 2012; Pecchenion & Pollard, 1997),

and that rates are unattractive due to the issue of adverse selection (Abel, 1985; Finkelstein

& Poterba, 2004; Mitchell & McCarthy, 2002). Still others argue that individuals self-insure

using other assets or resources (Laferrère, 2012; Vidal-Meliá & Lejárraga-Garcı́a, 2006).

However, the most common explanation for the annuitization puzzle is the bequest

motive. Several authors conclude that the bequest motive could be a barrier that might

reduce the demand for a life annuity (Bernheim, 1991; Friedman & Warshawsky, 1990;

Lockwood, 2012, 2018; Purcal & Piggott, 2008; Yaari, 1965). This is because annuitized

wealth is not bequeathable (Lockwood, 2018). Annuity benefits end at the death of the

annuitant.

2.2.1. The economics of annuities as bequest protection

As Yaari (1965) pointed out, although a bequest motive might prevent annuitization of all

assets, it should not prevent annuitization of any assets, that is, partial annuitization.

Longevity risk can result in all assets being completely exhausted due to living expenses

incurred during an exceptionally long life. This would result in the elimination of any

intended bequest. Annuitization can protect against this risk. As Davidoff et al. (2005, 1589)

explain, “partial annuitization can reduce the variation in the bequest.” Thus, partial annuiti-

zation should be attractive to people who have a bequest motive (Davidoff et al., 2005).

Those with a bequest motive should hold a portfolio which combines annuity wealth and

bequeathable wealth to maximize the marginal utility of both bequests and consumption

(Yaari, 1965).

Although theoretically sound, this concept may be sufficiently complex such as to require

explanation for a lay audience. Financial advisors may be able to bridge this complexity gap

between economic theory and practical advice by framing the annuity as a way to protect

the heirs’ inheritance.

Y. Yan and R. N. James / Financial Services Review 29 (2021) 277–291 279



2.2.2. Framing in behavioral finance

Frame dependence references circumstances in which people tend to make different choices

as the result of changes in phrasing or comparison salience, even though the relevant objective

facts remain the same (Baker & Nofsinger, 2010). In behavioral finance, examples of such

effects include gain/loss framing (Kahneman, Knetsch, & Thaler, 1991), narrow/broad framing

(Barberis, Huang, & Thaler, 2006), mental accounting (Thaler, 1985), and many others (Baker

& Nofsinger, 2010). Past experiments have shown that framing can influence decisions in

retirement planning in general (Choi et al., 2004) and annuity decisions in particular (Agnew et

al., 2008; Salisbury & Nenkov, 2016). For example, Salisbury and Nenkov (2016) found that

people were more interested in purchasing annuities described with life words than identical

annuities described with “death” words. Agnew et al. (2008) found that annuities became more

attractive when alternative investments were framed negatively. After reviewing evidence of

the annuity puzzle, Brown (2007) recommends that future annuity research should focus more

on such behavioral explanations.

2.2.3. The psychology of annuities as bequest protection

Another argument for the potential power of bequest protection framing for annuities comes

from past research connecting annuities and personal mortality salience. Terror Management

Theory, as well as a parallel utility maximization economic model (James, 2016), predict that

personal mortality salience (such as that triggered by death reminders) will generate two

responses: avoidance and pursuit of lasting social impact (a.k.a., symbolic immortality;

Pyszczynski, Greenberg & Solomon, 1999). Salisbury and Nenkov (2016) identified that annu-

ities, which can be thought of as a bet on one’s own mortality timing, do indeed serve as a death

reminder that triggers mortality salience.

The avoidance response will tend to reduce interest in annuities, simply because they are a

death reminder. Fitting with this response, Salisbury and Nenkov (2016) found that people were

more interested in purchasing annuities described as paying “if the annuity holder lives up to dif-

ferent ages” (life framing) than in purchasing annuities described as paying “depending on the

age when the annuity holder dies” (death framing; Salisbury & Nenkov, 2016, p. 423).

However, the resistance to annuities goes beyond simple avoidance of a death reminder.

The second response to mortality reminders predicted by theory is pursuit of lasting social

impact (a.k.a., “symbolic immortality”). This is the idea that although a person will die,

some part of them—their name, family, community, values, and so forth—will live on. A

bequest motive can be the financial expression of this psychological desire. Thus, annuities

serve as a death reminder, increasing the motivation to leave a bequest, but simultaneously

offer a financial benefit that disappears at death.

Confirming this idea, Williams and James (2019) found that higher levels of personal

mortality reminders generate a greater preference for an annuity paying lower income but

with a bequest provision—a form of life insurance. Such product combinations (life annuity

plus survivor bequest benefits) “solve” the desire for bequests. And indeed, Lockwood

(2012) estimates that about three-fourths of commercial annuities owned by recent retirees

have some provision that passes money to heirs after death.

However, these product combinations may not be optimal for financial outcomes. The

need to protect against longer than expected life span (annuity benefits) need not be

280 Y. Yan and R. N. James / Financial Services Review 29 (2021) 277–291



accompanied by a need to protect against shorter than expected life span (life insurance ben-

efits). The current experiment tests an alternative approach. Rather than coupling the annuity

with financial products offering actual bequest benefits, partial annuitization is simply ver-

bally framed as a form of bequest protection. If effective, this approach would provide the

opportunity to communicate the benefits of a life annuity in a way that matches underlying

psychological mechanisms, but without the need to combine it with other, potentially

unnecessary, financial products.

2.3. Hypotheses

Hypothesis 1: Framing an annuity as a protection for bequest goals will increase individuals’
interest in using an annuity in retirement.

Hypothesis 1A: Framing an annuity as a protection for bequest goals will increases individuals’
interest in using an annuity in retirement to a relatively greater extent for those who
were relatively more focused on leaving wealth to heirs (i.e., who had expressed a
stronger bequest motive).

Hypothesis 1 predicts that framing partial annuitization as protecting an inheritance for

heirs will increase interest in using an annuity in retirement. Hypothesis 1A predicts that the

effect of this framing will be more powerful on those who have expressed a relatively

greater bequest motive. Those who have expressed a relatively greater focus on leaving

wealth to heirs should logically be particularly influenced by this framing. Finding this to be

true would also lend support to the idea that increased interest resulting from framing the

annuity as bequest protection was arising through the proposed mechanisms, rather than, for

example, simply through generalized experimenter demand (Zizzo, 2010). In other words,

the effect of the framing would not be due solely to respondents changing their response to

match what they believe to be the answer desired by the experimenters, but would actually

relate to the respondent’s underlying bequest motive.

3. Data and results

3.1. Sample

Participants were recruited from adults in the United States via Amazon’s Mechanical Turk

(MTurk) platform with the headline description “University survey of opinions on retirement

planning.” Clicking on this generated the following description, “Survey of retirement planning

opinions. We are conducting an academic survey about opinions on retirement planning

options. This takes around 10 minutes (includes minimum timed responses and written text

requirements) and it is intended to advance research about people and their retirement plan-

ning, so please make sure you can commit the time. At the end of the survey, you will receive

a unique ‘completion code’ to paste into the box below to receive credit for taking our survey.”

There were 2,206 respondents who participated in the survey on October 18, 2019 with

2,160 responding to all questions used in the following analyses. The study was approved by

Y. Yan and R. N. James / Financial Services Review 29 (2021) 277–291 281



T
ab
le

1
K
ey

v
ar
ia
b
le
s
q
u
es
ti
o
n
te
x
t

A
n
n
u
it
y
ch
an
ce

(p
re

an
d
p
o
st
)

Im
ag
in
e
th
at
y
o
u
ar
e
6
5
y
ea
rs
o
ld

an
d
b
eg
in
n
in
g
re
ti
re
m
en
t.
Y
o
u
h
av
e
so
m
e
re
ti
re
m
en
t
sa
v
in
g
s
th
ro
u
g
h
y
o
u
r
em

p
lo
y
er

re
ti
re
m
en
t
p
la
n
an
d
ar
e
d
ec
id
in
g
h
o
w
to

m
an
ag
e
u
si
n
g
th
at
m
o
n
ey

in
th
e
co
m
in
g
y
ea
rs
.
Y
o
u
h
av
e
th
e
o
p
ti
o
n
to

p
u
t

y
o
u
r
re
ti
re
m
en
t
sa
v
in
g
s
in
to

an
an
n
u
it
y
th
at
w
il
l
g
iv
e
y
o
u
m
o
n
th
ly

p
ay
m
en
ts
ea
ch

y
ea
r
y
o
u
li
v
e.
W
h
en

y
o
u
p
u
t
y
o
u
r

sa
v
in
g
s
in
to

an
an
n
u
it
y
,
y
o
u
p
ay

a
lu
m
p
su
m

o
f
m
o
n
ey

u
p
fr
o
n
t.
In

re
tu
rn

fo
r
th
at
lu
m
p
-s
u
m

in
v
es
tm

en
t,
y
o
u
re
ce
iv
e
a

se
ri
es

o
f
re
g
u
la
r
m
o
n
th
ly

p
ay
m
en
ts
ea
ch

y
ea
r
y
o
u
li
v
e.

P
le
as
e
ra
te
th
e
p
er
ce
n
ta
g
e
li
k
el
ih
o
o
d
th
at
y
o
u
m
ig
h
t
co
n
si
d
er

p
u
rc
h
as
in
g
an

an
n
u
it
y
as

p
ar
t
o
f
y
o
u
r
re
ti
re
m
en
t
p
la
n
n
in
g
.

L
ik
el
ih
o
o
d
o
f
p
u
rc
h
as
in
g
an

an
n
u
it
y
:
[0
–
1
0
0
sl
id
er

b
ar
]

A
g
re
e
fr
am

in
g
ar
g
u
m
en
t

(y
es

=
1
)

O
n
e
b
en
efi
t
to

an
n
u
it
ie
s
is
th
at
th
ey

ca
n
p
ro
te
ct
an

in
h
er
it
an
ce

fo
r
h
ei
rs
.
N
o
m
at
te
r
h
o
w
lo
n
g
y
o
u
li
v
e,
y
o
u
w
il
l
h
av
e
li
fe
ti
m
e

in
co
m
e,
so

y
o
u
w
o
n
’t
b
e
fo
rc
ed

to
sp
en
d
aw

ay
y
o
u
r
o
th
er

in
v
es
tm

en
ts
o
r
as
se
ts
ju
st
to

h
av
e
re
g
u
la
r
in
co
m
e.

D
o
es

it
m
ak
e
se
n
se

to
y
o
u
h
o
w
h
av
in
g
g
u
ar
an
te
ed

li
fe
ti
m
e
in
co
m
e
fr
o
m

an
an
n
u
it
y
ca
n
p
ro
v
id
e
p
ro
te
ct
io
n
o
f
an

in
h
er
it
an
ce

fo
r
y
o
u
r
h
ei
rs
?

_
_
Y
es
,
th
at
m
ak
es

se
n
se

_
_
N
o
,
th
at
d
o
es
n
’t
m
ak
e
se
n
se

_
_
I’
m

n
o
t
re
al
ly

su
re

B
eq
u
es
t
m
o
ti
v
e

P
re
ce
d
in
g
q
u
es
ti
o
n
:
S
u
p
p
o
se

y
o
u
w
er
e
ag
e
6
5
an
d
h
ad

sa
v
ed

$
1
m
il
li
o
n
to

u
se

d
u
ri
n
g
re
ti
re
m
en
t
in

ad
d
it
io
n
to

so
ci
al
se
cu
ri
ty
.

Y
o
u
r
in
v
es
tm

en
ts
w
il
l
re
tu
rn

an
in
fl
at
io
n
-a
d
ju
st
ed

am
o
u
n
t
o
f
5
%

(e
.g
.,
$
5
0
,0
0
0
fr
o
m

$
1
m
il
li
o
n
)
p
er

y
ea
r.
W
h
ic
h
o
f
th
e

fo
ll
o
w
in
g
sp
en
d
in
g
p
la
n
s
w
o
u
ld

y
o
u
p
re
fe
r?

—
(p
ag
e
b
re
ak
)
—

O
n
a
sc
al
e
fr
o
m

0
to

1
0
0
,
h
o
w
m
u
ch

d
id

y
o
u
co
n
si
d
er

y
o
u
r
d
es
ir
e
to

le
av
e
m
o
n
ey

to
y
o
u
r
h
ei
rs
w
h
en

m
ak
in
g
th
e
sp
en
d
in
g

ch
o
ic
e
o
n
th
e
la
st
q
u
es
ti
o
n
?

C
o
n
si
d
er
ed

le
av
in
g
an

in
h
er
it
an
ce
:
[0
–
1
0
0
sl
id
er

b
ar
]

M
o
rt
al
it
y
sa
li
en
ce

T
o
w
h
at
ex
te
n
t
h
av
e
y
o
u
b
ee
n
th
in
k
in
g
ab
o
u
t
d
ea
th

in
th
e
p
as
t
se
v
er
al
m
in
u
te
s?

N
ev
er

[1
]
V
er
y
R
ar
el
y
[2
]
R
ar
el
y
[3
]
O
cc
as
io
n
al
ly

[4
]
F
re
q
u
en
tl
y
[5
]
V
er
y
F
re
q
u
en
tl
y
[6
]

P
le
as
e
ra
te
y
o
u
r
le
v
el
o
f
ag
re
em

en
t
w
it
h
th
e
fo
ll
o
w
in
g
st
at
em

en
t:

T
h
e
ta
sk
s
in

th
is
su
rv
ey

re
m
in
d
ed

m
e
o
f
d
ea
th
.

V
er
y
S
tr
o
n
g
ly

A
g
re
e
[6
]
S
tr
o
n
g
ly

A
g
re
e
[5
]
A
g
re
e
[4
]
D
is
ag
re
e
[3
]
S
tr
o
n
g
ly

D
is
ag
re
e
[2
]
V
er
y
S
tr
o
n
g
ly

D
is
ag
re
e
[1
]

T
o
w
h
at
ex
te
n
t
d
id

th
e
ta
sk
s
in

th
is
su
rv
ey

ev
o
k
e
th
o
u
g
h
ts
o
f
d
ea
th
?
N
ev
er

[1
]
V
er
y
R
ar
el
y
[2
]
R
ar
el
y
[3
]
O
cc
as
io
n
al
ly

[4
]

F
re
q
u
en
tl
y
[5
]
V
er
y
F
re
q
u
en
tl
y
[6
]

C
h
an
ce

li
v
e
to

1
0
0

O
n
a
sc
al
e
fr
o
m

0
%

to
1
0
0
%
,
w
h
at
ar
e
th
e
ch
an
ce
s
th
at
y
o
u
w
il
l
li
v
e
to

ag
e
1
0
0
o
r
o
ld
er
?

P
ro
b
ab
il
it
y
o
f
li
v
in
g
to

1
0
0
:
[0
–
1
0
0
sl
id
er

b
ar
]

282 Y. Yan and R. N. James / Financial Services Review 29 (2021) 277–291



the institutional review board for human research protection program of the second author’s

affiliated institution.

3.2. Variables

Table 1 reports the question text used to generate the key variables used in the following

analyses. Reports of education level were converted to number of years of education as fol-

lows, nine for “less than high school diploma,” 12 for “high school diploma,” 14 for “some

college or associates’ degree,” 16 for “bachelor’s degree,” and 19 for “graduate degree.”

Categorical reports of income were converted to $10,000 (for less than $10,000 category),

category midpoints of $15,000, $25,000, $35,000, $45,000, $55,000, $65,000, $75,000,

$85,000, $95,000, and $125,000, and finally $150,000 for greater than $150,000.

In the following regressions, socio-demographic control variables were also included.

More years of education may be associated with greater financial literacy (Van Rooij,

Lusardi, & Alessie, 2011). This might increase the likelihood of understanding the

bequest protection argument, but also increase the likelihood of having already consid-

ered the argument before the framing intervention. Alternatively, high education may be

associated with disagreement with the framing claim resulting from acceptance of an al-

ternative investment strategy, such as replacing annuities with mutual fund investments

and protecting an inheritance by accumulating sufficient wealth to provide for any

income needs (Pang & Warshawsky, 2010). Income is also associated with financial

sophistication and may impact the acceptance of the framing argument (Van Rooij,

Lusardi, & Alessie, 2011).

The longer one expects to live, the more relevant the bequest protection argument

becomes. A standard annuity involves a bet on one’s longevity. Thus, a simple mea-

surement of subjectively anticipated longevity—the chance one will live to age 100—is

included. Age is included as it has been associated with both financial literacy and the

confidence in one’s beliefs about financial products (such as annuities) and may impact

the influence of a framing argument (Finke, Howe, & Huston, 2017). Gender is

Table 2 Descriptive statistics

Variable Mean (SD) Min Max

Age 44.973 (16.061) 19 85
Male 0.509 (0.500) 0 1
Income 54.912 (34.108) 10 150
Education years 15.525 (1.995) 9 19
Annuity chance (Pre-Framing) 49.635 (27.179) 0 100
Annuity chance (Post-Framing) 53.992 (27.685) 0 100
Annuity interest increased 0.493 (0.500) 0 1
Annuity interest decreased 0.231 (0.421) 0 1
Agree framing argument 0.789 (0.408) 0 1
Bequest motive 54.390 (34.571) 0 100
Chance live to 100 31.010 (29.007) 0 100
Mortality salience 11.460 (3.525) 3 18
n 2,160

Y. Yan and R. N. James / Financial Services Review 29 (2021) 277–291 283



included as some research has found females to be more risk averse than males when

they make financial decisions in general (Jianakoplos & Bernasek, 1998) and with

regard to annuities in particular (Agnew et al., 2008). As such, females may respond

more favorably to the framing of an annuity as a hedge against a risk of total bequest

depletion.

Table 2 reports sample means, standard deviations, minimums, and maximums. The aver-

age respondent age was 45 with 51% of respondents being male. Average years of education

was 15.5, and average income was $55,000 per year.

3.3. Hypothesis 1: Results

Table 1 indicates that the mean percentage likelihood of purchasing an annuity as part

of retirement planning (Annuity Chance) increased following the bequest protection fram-

ing statement. A two-tailed paired t test finds that this difference is statistically significant

at p < .0001. Additionally, the percentage likelihood of purchasing an annuity as part of

retirement planning increased following the bequest protection framing statement for

49.3% of respondents but decreased for only 23.1% of respondents. Both findings support

Hypothesis 1.

These two outcomes measure the change in likelihood of using an annuity in two different

ways. The pre versus post comparison of probability changes measures the magnitude of the

change. This is important but is also subject to the effects of outliers where very large

changes take place. However, using a dummy variable for any increase in probability is not

subject to these outlier effects, as all positive changes are measured as a “1” regardless of

magnitude. Thus, it is useful to look at both forms of outcome measurements and this will

be repeated in the following regression analyses.

3.4. Hypothesis 1A: Results

Table 3 reports results for an ordinary least squares regression where the dependent vari-

able is the percentage likelihood of purchasing an annuity as part of retirement planning

Table 3 OLS regression with dependent variable: Annuity chance (Post-Framing)

Variable Parameter estimate (SE) p-Value

Intercept 10.793 (2.8989) 0.0002
Annuity chance (Pre-Framing) 1.0902 (0.0376) <0.0001
Annuity chance (Pre-Framing) SQ �0.0026 (0.0004) <0.0001
Bequest motive 0.0560 (0.0090) <0.0001
Mortality salience 0.1453 (0.0869) 0.0948
Male �1.9764 (0.6082) 0.0012
Income �0.0160 (0.0093) 0.0867
Education years �0.2583 (0.1597) 0.1060
Age �0.0432 (0.0192) 0.0241
Chance live to 100 0.0155 (0.0109) 0.1536

Note: n = 2,160; R2 = 0.7505. OLS = ordinary least squares.

284 Y. Yan and R. N. James / Financial Services Review 29 (2021) 277–291



following exposure to the bequest protection framing argument. The regression controls for

the percentage likelihood reported before the bequest protection framing and reflects the

change (pre vs. post) in this reported percentage likelihood. In addition, the regression con-

trols for the square of the percentage likelihood reported before the bequest protection fram-

ing. This is included because the dependent variable, Annuity Chance (Post-Framing), is

constrained to be between 0 and 100. As the initial percentage, a.k.a. Annuity Chance (Pre-

Framing), increases, the possible magnitude of an increase resulting from the framing falls,

suggesting a nonlinear relationship (e.g., if the likelihood of using an annuity was 99%

before the framing, there is little opportunity for increasing this likelihood). The results con-

firm this nonlinear relationship as both the linear and squared term are highly significant.

The increase resulting from the bequest protection framing argument is greater for those

who had expressed a greater bequest motive at p < .0001. This result supports the second hy-

pothesis (1A). An additional possible relationship is that the bequest protection framing

might be more powerful for those experiencing greater mortality salience. Terror manage-

ment theory suggests that those experiencing greater mortality salience will have an

increased interest in making a lasting social impact. Bequest protection framing should

appeal to this increased interest by showing the potential benefits of partial annuitization for

heirs. Contemplation of annuities has been found in past research to generate mortality sali-

ence (Salisbury & Nenkov, 2016), making this potential relationship particularly relevant for

the current experiment. Additionally, bequest motive was measured before the annuity ques-

tions while mortality salience was measured after the annuity questions. Thus, the level of

experienced mortality salience, controlling for bequest motive, would be sensitive to

changes generated by the annuity questions themselves. Table 3 shows a weak positive rela-

tionship between mortality salience and the increase in annuity interest following the

bequest protection framing intervention. However, this relationship is significant only at

p < .10.

Table 4 reports results from a probit analysis measuring whether or not the probability of

purchasing an annuity at retirement increased following the bequest protection framing argu-

ment. Unlike the analysis reported in Table 3 this analysis is not affected by large outliers,

such as where the reported percentage changed from 0 to 100 or vice-versa. Nevertheless,

the coefficient for bequest motive is still positive and significant at p < .0001. This confirms

Table 4 Probit analysis with dependent variable: Annuity chance increased

Parameter estimate
(SE)

Pr > v2

Intercept 0.0348 (0.2497) 0.8892
Bequest motive 0.0037 (0.0008) <0.0001
Mortality salience 0.0109 (0.0079) 0.1654
Male �0.1157 (0.0553) 0.0363
Income �0.0008 (0.0009) 0.3208
Education years �0.0012 (0.0145) 0.9316
Age �0.0061 (0.0017) 0.0004
Chance live to 100 0.0006 (0.001) 0.5650

Note: n = 2,160.

Y. Yan and R. N. James / Financial Services Review 29 (2021) 277–291 285



that the relationship identified in Table 3 is not driven only by large outliers but reflects a

general tendency among the respondents. Thus, the likelihood that interest in an annuity will

increase to some extent following the bequest protection framing also increases as bequest

motive increases. This result further supports the second hypothesis (1A).

The direction and significance of the relationship with age and gender remains consistent

with the results reported in Table 3. However, the relationship with mortality salience

becomes nonsignificant (p = .1654).

The primary interest in this analysis was to explore the results from introducing a bequest

protection framing argument on interest in purchasing an annuity as part of retirement plan-

ning. Bequest protection framing may fail to have an effect because respondents disagree

with the bequest protection argument. However, it may also fail to have an effect even if

respondents agree with the bequest protection argument. This could result from respondents

having already considered such a justification before its introduction in the framing argu-

ment or because other unmeasured factors make such a justification irrelevant for the re-

spondent. Thus, Table 5 reports results from an additional measurement of the effectiveness

of this framing argument: whether respondents agree that the argument itself makes sense.

This measurement of the power of the bequest protection framing argument also shows a

positive association with bequest motive (p < .0001). Thus, those with a higher bequest

motive are more likely to agree that the bequest protection framing argument makes sense.

One difference as compared with the previous outcome measurement is that mortality

salience is a highly significant predictor (p < .0001) of agreeing with the bequest protection

framing argument. This may be because the increased desire for making a lasting impact

resulting from mortality salience may increase acceptance of the argument describing such

an impact. Future investigation of this relationship may be critical as previous research

shows contemplating annuity purchases increases mortality salience (Salisbury & Nenkov,

2016). Thus, a framing argument that works particularly well in the presence of mortality

salience may be particularly effective in a real-world application. However, in the present

experiment this relationship may arise because acceptance and understanding of the bequest

protection argument actually generates mortality salience. Future experimental research may

be able to disentangle this relationship more precisely.

Table 5 Probit analysis with dependent variable: Agree with framing argument

Parameter estimate
(standard error)

Pr > v2

Intercept 0.7593 (0.2819) 0.0071
Bequest motive 0.0038 (0.0009) <0.0001
Mortality salience 0.0355 (0.0090) <0.0001
Male �0.0054 (0.0627) 0.9311
Income �0.0001 (0.0009) 0.8822
Education years �0.0262 (0.0163) 0.1085
Age �0.0054 (0.0019) 0.0041
Chance live to 100 0.0036 (0.0011) 0.0014

Note: n = 2,160.

286 Y. Yan and R. N. James / Financial Services Review 29 (2021) 277–291



4. Discussion

This research examines bequest motives and annuitization in a new way. It examines an

intervention framing an annuity as a protection for inheritance goals from longevity risk. In

the presence of other assets intended to be left as a bequest, annuitization can be viewed as

protecting those assets against consumption due to a person’s excessive longevity. Thus, in

the absence of (partial) annuitization, these assets might be completely exhausted due to liv-

ing expenses incurred during an exceptionally long life. This bequest protection frame

presents the annuity as a downside protection for inheritance goals.

We hypothesized that this framing intervention would encourage interest in using

annuities in retirement. The results are consistent with this hypotheses that framing mat-

ters. On average, introducing the bequest protection framing intervention increases inter-

est in using an annuity in retirement. Additionally, regression models showed that a

greater bequest motive leads to a more positive effect of the bequest protection framing

intervention on the predicted likelihood of using an annuity in retirement. Thus, the

impact of the intervention is likely not due exclusively to generalized experimenter

demand effects. This result matches economic arguments for partial annuitization as ben-

efitting heirs by reducing bequest volatility but suggests a necessary role for the advisors

to explain this justification.

This bequest protection framing may represent an additional, practical approach to

addressing the annuitization puzzle (Bernheim, 1991; Bütler & Teppa, 2007; Yaari, 1965).

This preliminary evidence suggests a potential method to directly address previous findings

of a negative association between bequest motives and annuitization (Bernheim, 1991;

Bütler & Teppa, 2007; Yaari, 1965). Without special framing, this negative association is

obvious. Annuity wealth is not bequeathable. Therefore, people who have a bequest motive

may be less likely to purchase life annuities. However, bequest protection framing appears

to most strongly influence those who have higher bequest motives.

Developing a method, such as that tested here, to address these bequest motive objections

may be particularly important because the act of annuity contemplation may actually

increase bequest motives. Past research has shown that annuity contemplation generates

mortality salience (Salisbury & Nenkov, 2016) and also that mortality salience generates an

increased desire for lasting social impact, a.k.a., pursuit of symbolic immortality (James,

2016; Pyszczynski, Greenberg, & Solomon, 1999). This results in the outcome, demon-

strated by Williams and James (2019), that increasing mortality salience increases interest in

annuities that pay less income but include bequest benefits relative to those that pay more

income with no such bequest benefit. However, this outcome is unlikely to be an ideal finan-

cial planning choice. The current reframing approach offers the advantage that no product

changes are needed to increase the acceptability of the annuity choice.

Of course, the present experiment measures responses to a hypothetical question about an

annuity purchase decision to be made at age 65. It does not measure actual purchase behav-

ior. Behavioral intentions can differ from actual behavior. However, because this analysis

focuses on the change in behavioral intentions, the results can be instructive even in the

presence of imprecise measurements or additional barriers between intentions and action.

Y. Yan and R. N. James / Financial Services Review 29 (2021) 277–291 287



4.1. Financial planning practice implications: Client conversations

Economic theory demonstrates that partial annuitization may be ideal even for those indi-

viduals with a strong bequest motive. Partial annuitization reduces the volatility of a bequest

by providing protection against reduction or exhaustion of the bequest assets due to income

needs resulting from longevity risk. The current results suggest that this argument, when pre-

sented in nontechnical terms, can be persuasive to the general public.

However, the argument is not uniformly attractive. It is more compelling for some people

than others. In particular, the argument is more compelling for those who begin with a

higher bequest motive. Additionally, this framing is more effective among women and those

who are younger.

Determining bequest motive need not be a complicated task for the advisor. Some clients

may bring up such concerns when annuities are discussed. In this study, bequest motives

were measured by the response to the question, “how much did you consider your desire to

leave money to your heirs when making the [retirement] spending choice on the last ques-

tion?” Similarly simple questions may also elicit the relevant motivations from clients.

Where such bequest motives exist, the bequest protection framing for annuities may be par-

ticularly effective.

4.2. Financial planning practice implications: Client annuity usage

The results presented here have two practical implications for client choice of annuity

products. First, the results can be helpful where annuities are an appropriate financial

option for clients. The results demonstrate the effectiveness (especially for certain peo-

ple) of a bequest protection framing argument. Second, the results can be helpful where

the typical form of annuities (annuities with survivor benefits) are less than optimal.

Where Williams and James (2019) showed that annuities producing less income and

greater bequest benefits become more attractive with increased mortality salience, this

research shows that it is not necessarily required to change the product to include more

bequest benefits, but it can work to simply change the product framing to highlight the

otherwise unrealized bequest benefits.

4.3. Annuity advantages

Finding a cost-free way to encourage the use of annuities—without the addition of poten-

tially unnecessary life-insurance-like bequest benefits—can be potentially beneficial for cli-

ents. The first reason is the risk. Compared with investing in the stock market, the risk of

investing in annuities is relatively low and the annuity lifetime income or certain period

income is guaranteed. For example, the sequence of returns risk, which could have a signifi-

cant impact on retirees’ portfolios, does not arise with annuities but does occur with inves-

ting in the stock market.

A Monte Carlo simulation shows the impact from this sequence of returns risk. The simu-
lation assumes a retiree with an initial investment of $500,000 dollars, average returns for

288 Y. Yan and R. N. James / Financial Services Review 29 (2021) 277–291



stock market of 6% (annually), a standard deviation of return of 20%, a time horizon for
investment of 20 years, and the amount added to the investment portfolio of $15,000 at the
end of each year. In 10,000 iterations, the ending portfolio (mean $2,192,524; median
$1,633,929; standard deviation $1,966,518) varies widely from the fifth percentile outcome
of $465,544 to the 95th percentile outcome of $5,722,992. If the investors suffers a “bad
time” at the beginning of the investment period, that is, negative returns at the beginning of
the investment period (e.g., caused by an economic crisis or other global issues), the invest-
ors might suffer a huge initial loss with regard to the investment portfolio. If retirees have
such a loss in their investments, their assets might be completely exhausted, and they may
not have sufficient wealth for their retirement. However, the returns for annuity investments
are stable and the annuity income is typically guaranteed. For example, in the above simula-
tion, placing 10% ($50,000) of the initial investment into a fixed twenty-year guaranteed
annuity with a lower (4%) return increases the fifth percentile outcome by 30% but decreases
the 95th percentile outcome by only 3%.

Compared with the stock and mutual fund market, the bond market is a closer alternative

to annuities for retirees. However, bond investors still have to face many risks, such as credit

risk (bonds may not pay as expected), inflation risk (as the inflation rate increases, the cou-

pon income values becomes less), reinvestment risk (investors may not be able to find a sim-

ilar bond investment option after existing bonds mature), and liquidity risk (bonds may be

unavailable to meet short-term financial goals). Thus, if retirees only invest in stock, mutual

funds, and bonds, they have greater risk for meeting their short-term and long-term financial

goals. Having an annuity could help retirees have lifetime income or certain period income,

which could be a protection for their financial goals and reduce risk (including systematic

risk from overall market returns and unsystematic risk from individual investment selection).

Therefore, annuities could improve retirees’ welfare and could be valuable for retirees’

portfolios.

4.4. Annuity disadvantages

Despite the previous advantages, the reality of annuity pricing may reduce the advan-

tages compared with theoretical potential.There is adverse selection with regard to the

mortality risk in the annuity market (Blake, 1999). This is the risk that only individuals

who think they will live longer than the average for the population based on their own

medical and family histories will choose to purchase annuities. And also, conversely,

those who know of a major mortality risk factor will not ever purchase an annuity (i.e.,

sick people do not buy annuities). However, insurance companies do not have the same

access to this information with the same degree of reliability as do annuity purchasers

(Blake, 1999). Therefore, there is asymmetric information between the annuity insurance

company and prospective annuitants. Insurance companies may not be able to perfectly

distinguish the prospective annuitants who have heavier mortality risk (insurance compa-

nies make money) from those who have lighter mortality risk (insurance companies lose

money). However, companies realize that people who voluntarily purchase an annuity

have a lighter mortality risk than the typical person of the same age. Thus, pricing for all

prospective annuitants will reflect this reality, potentially making the annuity (in

Y. Yan and R. N. James / Financial Services Review 29 (2021) 277–291 289



practice) a poor investment choice for those with a typical life expectancy. Once this is

combined with the need for paying commissions, advertising, and profits, the theoreti-

cally advantageous annuity may become practically less attractive.

References

Abel, A. (1985). Capital Accumulation and Uncertain Lifetimes With Adverse Selection. NBER Working Paper

Series, 1664. Cambridge, MA: National Bureau of Economic Research.

Agnew, J. R., Anderson, L. R., Gerlach, J. R., & Szykman, L. R. (2008). Who chooses annuities? An experi-

mental investigation of the role of gender, framing, and defaults. American Economic Review, 98, 418–

422.

Baker, H. K., & Nofsinger, J. R. (2010). Behavioral finance: An overview. In H. K. Baker & J. R. Nofsinger

(Eds.), Behavioral Finance: Investors, Corporations and Markets (pp. 3–20). Hoboken, NJ: John Wiley &

Sons.

Banerjee, S. (2013). Annuity and lump-sum decisions in defined benefit plans: The role of plan rules. EBRI Issue

Brief, 381, 1–20.

Barberis, N., Huang, M., & Thaler, R. H. (2006). Individual preferences, monetary gambles, and stock market

participation: A case for narrow framing. American Economic Review, 96, 1069–1090.

Bernheim, B. D. (1991). How strong are bequest motives? Evidence based on estimates of the demand for life in-

surance and annuities. Journal of Political Economy, 99, 899–927.

Benartzi, S., Previtero, A., & Thaler, R. H. (2011). Annuitization puzzles. Journal of Economic Perspectives, 25,

143–164.

Blake, D. (1999). Annuity markets: Problems and solutions. The Geneva Papers on Risk and Insurance - Issues

and Practice, 24, 358–375.

Brown, J. R., & Poterba, J. M. (2000). Joint life annuities and annuity demand by married couples. The Journal

of Risk and Insurance, 67, 527–553.

Brown, J. R. (2001). Private pensions, mortality risk, and the decision to annuitize. Journal of Public Economics,

82, 29–62.

Brown, J. R. (2003). Redistribution and insurance: Mandatory annuitization with mortality heterogeneity.

Journal of Risk & Insurance, 70, 17–41.

Brown, J. R. (2007). Rational and Behavioral Perspectives on the Role of Annuities in Retirement Planning.

NBER Working Paper No. w13537. Cambridge, MA: National Bureau of Economic Research.

Bütler, M., & Teppa, F. (2007). The choice between an annuity and a lump sum: Results from Swiss pension

funds. Journal of Public Economics, 91, 1944–1966.

Cox, S. H., & Lin, Y. (2007). Natural hedging of life and annuity mortality risks. North American Actuarial

Journal, 11, 1–15.

Choi, J. J., Laibson, D., Madrian, B. C., & Metrick, A. (2004). For better or for worse: Default effects and 401

(k) savings behavior. In D. A. Wise (Ed.), Perspectives on the Economics of Aging (pp. 81–126). Chicago, IL:

University of Chicago Press.

Davidoff, T., Brown, J. R., & Diamond, P. A. (2005). Annuities and individual welfare. American Economic

Review, 95, 1573–1590.

Donnelly, C. (2015). Actuarial fairness and solidarity in pooled annuity funds. ASTIN Bulletin, 45, 49–74.

Finke, M. S., Howe, J. S., & Huston, S. J. (2017). Old age and the decline in financial literacy. Management

Science, 63, 213–230.

Finkelstein, A., & Poterba, J. (2004). Adverse selection in insurance markets: Policyholder evidence from the

UK annuity market. Journal of Political Economy, 112, 183–208.

Friedman, B., & Warshawsky, M. (1990). The cost of annuities: Implications for saving behavior and bequests.

The Quarterly Journal of Economics, 105, 135–154.

290 Y. Yan and R. N. James / Financial Services Review 29 (2021) 277–291



James, R. N., III., (2016). An economic model of mortality salience in personal financial decision making:

Applications to annuities, life insurance, charitable gifts, estate planning, conspicuous consumption, and

healthcare. Journal of Financial Therapy, 7, 62–82.

Jianakoplos, N. A., & Bernasek, A. (1998). Are women more risk averse? Economic Inquiry, 36, 620–630.

Kahneman, D., Knetsch, J. L., & Thaler, R. H. (1991). Anomalies: The endowment effect, loss aversion, and sta-

tus quo bias. Journal of Economic Perspectives, 5, 193–206.

Knell, M. (2018). Increasing life expectancy and NDC pension systems. Journal of Pension Economics &

Finance, 17, 170–199.

Laferrère, A. (2012). Housing wealth as self-insurance for long-term care. In Financing Long-Term Care in

Europe (pp. 73–90). London, UK: Palgrave Macmillan.

Lockwood, L. M. (2012). Bequest motives and the annuity puzzle. Review of Economic Dynamics, 15, 226–243.

Lockwood, L. M. (2018). Incidental bequests and the choice to self-insure late-life risks. American Economic

Review, 108, 2513–2550.

Mitchell, O. S., & McCarthy, D. (2002). Estimating international adverse selection in annuities. North American

Actuarial Journal, 6, 38–54.

Modigliani, F. (1986). Life cycle, individual thrift, and the wealth of nations. Science, 234, 704–712.

Mottola, G., & Utkus, S. (2008). Lump sum or annuity? An analysis of choice in DB pension payouts. Pension

Benefits, 17, 2–4.

Pang, G., & Warshawsky, M. (2010). Optimizing the equity-bond-annuity portfolio in retirement: The impact of

uncertain health expenses. Insurance: Mathematics and Economics, 46, 198–209.

Pecchenino, R., & Pollard, P. (1997). The effects of annuities, bequests, and aging in an overlapping generations

model of endogenous growth. The Economic Journal, 107, 26–46.

Peijnenburg, K., Nijman, T., & Werker, B. J. (2016). The annuity puzzle remains a puzzle. Journal of Economic

Dynamics and Control, 70, 18–35.

Purcal, S., & Piggott, J. (2008). Explaining low annuity demand: An optimal portfolio application to Japan.

Journal of Risk & Insurance, 75, 493–516.

Pyszczynski, T., Greenberg, J., & Solomon, S. (1999). A dual process model of defense against conscious and

unconscious death-related thoughts: An extension of terror management theory. Psychological Review, 106,

835–845.

Salisbury, L. C., & Nenkov, G. Y. (2016). Solving the annuity puzzle: The role of mortality salience in retirement

savings decumulation decisions. Journal of Consumer Psychology, 26, 417–425.

Scott, J. S. (2008). The longevity annuity: An annuity for everyone? Financial Analysts Journal, 64, 40–48.

Thaler, R. (1985). Mental accounting and consumer choice.Marketing Science, 4, 199–214.

Van Rooij, M., Lusardi, A., & Alessie, R. (2011). Financial literacy and stock market participation. Journal of

Financial Economics, 101, 449–472.

Vidal-Meliá, C., & Lejárraga-Garcı́a, A. (2006). Demand for life annuities from married couples with a bequest

motive. Journal of Pension Economics and Finance, 5, 197–229.

Williams, J. A., & James, R. N. , III., (2019). Bequest provision preferences in commercial annuities: An experi-

mental test of the role of mortality salience. Journal of Financial Counseling and Planning, 30, 121–131.

The World Bank. (2019). Life Expectancy At Birth, Total (Years). Available at https://data.worldbank.org/

indicator/SP.DYN.LE00.IN?cid=GPD_10

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

Studies, 32, 137–150.

Zizzo, D. J. (2010). Experimenter demand effects in economic experiments. Experimental Economics, 13, 75–98.

Y. Yan and R. N. James / Financial Services Review 29 (2021) 277–291 291