FINSER_31-2_Complete_Issue


Investment literacy, overconfidence and cryptocurrency
investment

Kyoung Tae Kima,*, Sherman D. Hannab, Sunwoo T. Leec

aDepartment of Consumer Sciences, University of Alabama, 316-C Adams Hall, Box 870158, Tuscaloosa,
AL 35487, USA

bDepartment of Human Sciences, The Ohio State University, 1787 Neil Ave, Columbus, OH 43210, USA
cSchool of Administrative Studies, York University, Atkinson Building, 252, 4700 Keele Street, Toronto,

ON M3J1P3, Canada

Abstract

Cryptocurrency has been increasingly popular with investors. Using the 2018 National Financial
Capability Study Investor survey, we examined the association between investment literacy and
cryptocurrency investment—about 13% of investors invested in cryptocurrency directly or indirectly.
Results from regression analyses show that objective investment literacy was negatively while sub-
jective literacy was positively associated with holding cryptocurrency. Overconfident investors were
more likely to invest in cryptocurrency, and results were robust across three overconfidence meas-
ures. This study has implications for investment advice, financial education, and research. © 2023
Academy of Financial Services. All rights reserved.

JEL classification: G5; G11; G53

Keywords: Investment literacy; Financial knowledge; Overconfidence; Investment; Cryptocurrency

1. Introduction

Cryptocurrency has become increasingly popular since 2008, since the invention of
Bitcoin by Satoshi Nakamoto (Nakamoto, 2008). Although Bitcoin still dominates the cryp-
tocurrency market, nearly 8,600 cryptocurrencies are currently trading around the world
with a combined market capitalization of US$1.48 trillion as of March 2, 2020 (Coin Market

*Corresponding author: Tel.: 205-348-9167, fax: 205-348-8721.
E-mail address: ktkim@ches.ua.edu

1057-0810/23/$ – see front matter © 2023 Academy of Financial Services. All rights reserved.

Financial Services Review 31 (2023) 121–132



Cap, 2020). The total value of holdings in cryptocurrencies is still small relative to holdings
of gold and financial investments but is being increasingly studied as a component of portfo-
lios of wealthy investors (Zhao, 2021). One indicator of the growing importance of crypto-
currencies is concern about the impact on the environment of the production of
cryptocurrencies (Sorkin, 2021).

Although investment in cryptocurrency has been rapidly growing, there are not many
studies on factors related to people investing in cryptocurrency. One of the major reasons is
the limited availability of appropriate survey data and questionnaires. The 2018 National
Financial Capability Study Investor Survey (thereafter, 2018 NFCS Investor Survey)
released by FINRA Investor Education Foundation collects a series of questions about cryp-
tocurrency investment. This allows us a unique opportunity to identify factors associated
with investing in cryptocurrency by US investors. Given limited knowledge about investing
in cryptocurrencies and the extreme volatility as an investment, what type of investor holds
cryptocurrency investments?

The purpose of this study is to investigate the association between investment literacy and
cryptocurrency investment. We tested investment literacy in three ways: (1) objective liter-
acy, (2) subjective literacy, and (3) overconfidence in investment literacy. To check the
robustness of our results, we used three different indicators of overconfidence in investment
literacy, measured by divergence between objective and subjective literacy. In addition to
the analysis on the association between investment literacy and investing in cryptocurrency,
we examined various socioeconomic factors related to investing in cryptocurrency. This
study contributes to the existing literature in two ways. While there has been increasing
attention to cryptocurrency investment, our study is one of first attempts to explore the asso-
ciation between investment literacy overconfidence and investing in cryptocurrency of US
investors. Further, we used an index of correct answers to investment knowledge questions,
which captures one’s investment literacy more comprehensively than commonly used meas-
ures of financial literacy (Lusardi & Mitchell, 2014). The major findings of this study pro-
vide important insights for investment advisors, financial educators, and researchers.

2. Literature review and theoretical consideration

2.1. Previous studies on cryptocurrency investments

While cryptocurrency is labeled “currency,” the actual adoption of cryptocurrency as a
payment method is still a topic of discussion. Cryptocurrency has advantages such as confi-
dentiality, reliability of information transmission, and flexible transactions, but drawbacks
such as significant volatility, lack of organized platforms, and the difficulty of projection
hinder it from being integrated to the global financial market (Tasca et al., 2018; Titov et al.,
2021).

Cryptocurrency has also gained a tremendous amount of attention as an investment vehi-
cle. Chuen et al. (2017) argued that due to low correlation between cryptocurrencies and tra-
ditional assets as well as the higher daily expected return of the former option, investing in

122 K. T. Kim et al. / Financial Services Review 31 (2023) 121–132



cryptocurrencies could help one diversify their portfolio risks. When examined as a form of
financial asset, cryptocurrency presented quite similar dynamics to stock investments (Liang
et al., 2019).

Previous studies have documented several factors related to cryptocurrency investments.
Socio-demographic factors such as age, gender, and education were correlated to cryptocur-
rency investment (Ante et al., 2022; Hasso et al., 2019; Henry et al., 2018). Since cryptocur-
rency is based on a cutting-edge technological innovation, many studies have also used the
Technology Acceptance Model as a theoretical background for cryptocurrency investment.
Studies confirmed that the new technology’s perceived usefulness and ease of use as well as
the risk associated with it affect one’s intention to utilize cryptocurrency (Arias-Oliva et al.,
2019; Bharadwaj & Deka, 2021).

One phenomenon that might be related to the growth of cryptocurrency investment is the
bandwagon effect, or herding behavior (Avital et al., 2016; Bouri et al., 2019; da Gama
Silva et al., 2019). Herding behavior combined with the subsequent high volatility could
pose significant risks on the financial stability of the cryptocurrency investors. Thus, it calls
for a cautious approach to cryptocurrency investment and emphasizes the importance of
understanding whether investors selected cryptocurrency as an investment vehicle are capa-
ble of understanding complex dynamic and the potential consequences of their choices.

2.2. Financial literacy and investment decisions

Unlike what traditional economists have assumed and approached, many studies present
evidence that investors are functioning under bounded rationality (de Bondt et al., 2008).
The theory of bounded rationality (Simon, 2000) posits that individuals have limited ability
to assess given information and make optimal decisions. This is mainly due to the complex-
ity of environments, limited mental capacity, and limited resources (Ibrahim, 2009).
Financial literacy is a measure of individuals’ ability to understand situations and make opti-
mal financial decisions based on their assessments (Seay et al., 2017). Therefore, those with
higher financial literacy would behave in a more appropriate manner, and financial mistakes
would be more common among the financially illiterate people. Enhancing financial literacy
could help reduce the probability of a major systematic financial crisis due to bounded
rationality (Siriopoulos, 2021). Grounded within the theory of bounded rationality, previous
studies have documented the positive link between financial literacy and various financial
outcomes including stock ownership (Kimball & Shumway, 2006; van Rooij et al. 2011),
diversification (Goetzmann & Kumar, 2008; Guiso & Jappelli, 2008; Shin et al., 2020),
wealth accumulation and planning for retirement (Lusardi & Mitchell, 2007).

Cryptocurrency investment is one of the complex financial decisions which would require
one’s understanding of the concept and characteristics of cryptocurrency and relatively
unique procedures of investment, but there have been few studies conducted on the link
between cryptocurrency and financial literacy (Arias-Oliva et al., 2019; Zhao & Zhang,
2021). Financial literacy has been conceptualized as objective and subjective measures,
where objective financial literacy indicates one’s actual understanding of financial concepts
while subjective financial literacy refers to one’s perceived understanding of the matter

K. T. Kim et al. / Financial Services Review 31 (2023) 121–132 123



(Kim et al., 2020; Robb et al., 2015). Objective financial literacy has been found negatively
related to cryptocurrency use or acceptance (Zhao & Zhang, 2021) while subjective financial
literacy has been found positively related to (Gupta et al., 2020) or did not have a statisti-
cally significant relation to cryptocurrency utilization (Arias-Oliva et al., 2019). Based on
the findings from previous studies, we constructed the following two research hypotheses on
the association between investment literacy and cryptocurrency investment.

Hypothesis 1: Objective investment literacy is negatively associated with investing in crypto-
currency.

Hypothesis 2: Subjective investment literacy is positively associated with investing in crypto-
currency.

We focused on the link between overconfidence in investment literacy and cryptocurrency
investments, to provide new insights into factors related to holding cryptocurrency as an
investment. Investors may be influenced by herd behavior instead of making investment
decisions based on the value of cryptocurrencies; therefore, it is useful to investigate the
cryptocurrency investment from the perspective of bounded rationality, and we extended it
to the aspect of overconfidence in this study.

Overconfidence bias has been described as (1) overprecision or (2) miscalibration, indi-
cating a systematic overweighting of the accuracy of one’s own literacy (e.g., Robb et al.
2015; Xia et al., 2014). Researchers have examined the relationship between overconfidence
and irrational behaviors such as excessive trading, excessive risk taking (Abreu & Mendes,
2012; Barber & Odean, 2000), under diversification (Chu et al. 2017; Shin et al., 2020; Xia
et al., 2014) and high-cost borrowing (Robb et al., 2015). Some studies have raised concerns
about the divergence between objective and subjective financial knowledge (e.g., Kim et al.,
2020; Robb et al., 2015). False confidence in one’s ability to understand complex investment
concepts would lead to suboptimal financial decisions.

In this study, we propose the following research hypothesis on the association between
investment literacy overconfidence and cryptocurrency investment.

Hypothesis 3: Overconfidence in investment literacy is positively associated with investing in
cryptocurrency.

3. Method

3.1. Dataset and analytic sample

This study utilized data from the 2018 National Financial Capability Study (NFCS) and
the follow-up 2018 Investor Survey, which includes a subset of respondents who had indi-
cated ownership of nonretirement investments. The 2018 NFCS state-by-state data collected
approximately 500 observations per state plus the District of Columbia, which leads to the
total sample of 27,091 adults in the United States, with approximately 500 observations per
state plus the District of Columbia. The 2018 NFCS Investor Survey includes a sample
drawn from individuals who indicated owning nonretirement investments. To explore the
associations between investment literacy and investing in cryptocurrency, the analytic

124 K. T. Kim et al. / Financial Services Review 31 (2023) 121–132



sample started with the 2,003 individuals who completed both the 2018 NFCS and the 2018
NFCS Investor Survey. The final sample includes 1,819 investors, excluding missing
responses.

3.2. Dependent variables

The 2018 NFCS Investor Survey includes questions on cryptocurrency. The dependent
variable is a binary indicator whether the respondent has invested in cryptocurrency, based
on the following question, “Have you invested in cryptocurrencies, either directly or through
a fund that invests in cryptocurrencies?”

3.3. Investment literacy and overconfidence

The 2018 NFCS Survey includes questions designed to measure investment literacy cov-
ering various investment-related topics and concepts. We created an objective investment
literacy index, measured as each respondent’s number of correct answers to 10 questions
and ranged from 0 to 10. The topics of investment literacy cover stock, bond, bankruptcy,
investment risk, investment return, municipal bond, stock margin, selling short, investment
indicator, and index fund. The subjective assessment of investment knowledge was meas-
ured by the following question: “On a scale from 1 to 7, where 1 means very low and 7
means very high, how would you assess your overall knowledge about investing?”

Following the approach of Kim et al. (2020), we tested three measures of investment liter-
acy overconfidence: (1) overconfidence defined as having higher than the sample median for
subjective literacy, but lower than the sample median for objective literacy; (2) a continuous
measure of divergence between objective and subjective financial literacy; and (3) the resid-
ual from a least squares regression of objective literacy on subjective literacy.

3.4. Control variables

In addition to investment literacy variables, we included age, gender (male, female), mari-
tal status (married, single, separated/divorced/widow(er)), having a dependent child, race/
ethnicity (White, Black, Hispanic, Asian/others), employment status (full-time employee,
self-employed, part-time employee, homemaker, student, disabled, unemployed, retired),
education (high school diploma or lower, some college, associate degree, bachelor degree,
postbachelor degree), household income, investment assets in nonretirement account, and
region (state of residence) as control variables.

3.5. Empirical specification

We used logistic regression models to investigate the effects of various factors associated
with an investment in cryptocurrency, especially for the role of investment literacy and its
overconfidence measured by divergence between objective and subjective literacy. Both de-
scriptive and multivariate results were weighted using the sampling weight provided by
NFCS Investor survey. Empirical models are expressed as follows;

K. T. Kim et al. / Financial Services Review 31 (2023) 121–132 125



Pr Cð Þi = log
Pr Cð Þi

1# Pr Cð Þi
!

" #
= b 0 þ b 1ILi þ b 2Xi þ « i

Where, Pr Cð Þi is the probability of investing in cryptocurrency; b 0 is an intercept; IL
denotes the level of investment literacy including objective, subjective investment literacy
(baseline model) and three overconfidence induce. The control variables and state of resi-
dence are denoted as Xi.

4. Results

4.1. Descriptive results

As shown in Table 1, about 13% of investors invested in cryptocurrency directly or indi-
rectly. Based on the proportion of all households with nonretirement investments, we can
infer that less than 4% of all households held cryptocurrency. We also tested the mean dif-
ferences in investment literacy of cryptocurrency investors and noninvestors. Respondents
owning cryptocurrency investments had lower objective investment literacy scores and
higher subjective literacy than noncryptocurrency investors. In addition, the level of over-
confidence in investment literacy was higher for cryptocurrency owners than nonowners
across three overconfidence indexes. Socio-demographic characteristics of the analytic sam-
ple are presented in Appendix.

4.2. Multivariate results

Results of logistic regressions on cryptocurrency investment are reported in Table 2. In
Model A (baseline), objective investment literacy was negatively, but subjective literacy was
positively associated with investing in cryptocurrency. In particular, a one unit increase in
objective literacy decreased the odds of investing in cryptocurrency by 10.5%; while a one

Table 1 Descriptive results of selected variables, 2018 NFCS investor survey

Variables All investors Investing in
cryptocurrency

Not investing in
cryptocurrency

Dependent variable
Investing in cryptocurrency 12.90% — —

Investment literacy
Objective literacy (0–10), mean (median) 4.77 (5.00) 3.96 (4.00)*** 4.89 (5.00)
Subjective literacy (1-7), mean (median) 4.79 (5.00) 5.55 (6.00)*** 4.68 (5.00)

Overconfidence in investment literacy
Index 1 (high subjective/low objective literacy) 28.28% 58.7%*** 23.9%
Index 2 (numerical difference), mean (median) 0.02 (0.00) 1.58 (3.00)*** #0.21 (0.00)
Index 3 (residual measure), mean (median) 0.00 (0.18) 0.86 (1.04)*** #0.13 (0.04)

Notes. Weighted results. t test or x2 are conducted for a group comparison.
Significance level: *p < .05, **p < .01, ***p < .001.

126 K. T. Kim et al. / Financial Services Review 31 (2023) 121–132



T
ab
le

2
L
o
g
is
ti
c
re
g
re
ss
io
n
s
o
n
cr
y
p
to
cu
rr
en
cy

in
v
es
tm

en
t,
2
0
1
8
N
F
C
S
in
v
es
to
r
su
rv
ey

V
ar
ia
b
le
s

(A
)
B
as
el
in
e
m
o
d
el

(B
)
O
ve
rc
on
fi
de
nc
e

in
de
x
1

(h
ig
h
su
bj
ec
ti
ve
,

lo
w
ob
je
ct
iv
e)

(C
)
O
v
er
co
nfi

d
en
ce

in
d
ex

2
a

(d
is
cr
ep
an
cy
)

(D
)
O
v
er
co
n
fi
d
en
ce

in
d
ex

3b

(r
es
id
u
al
)

C
o
ef
f.

S
E

C
o
ef
f.

S
E

C
o
ef
f.

S
E

C
o
ef
f.

S
E

In
v
es
tm

en
t
li
te
ra
cy

v
ar
ia
b
le
s

O
b
je
ct
iv
e
li
te
ra
cy

#
0
.1
1
1
4
*

0
.0
4
7
2

—
—

—
—

—
—

S
u
b
je
ct
iv
e
li
te
ra
cy

0
.5
9
1
*
*
*

0
.0
8
1
2

—
—

—
—

—
—

O
v
er
co
n
fi
d
en
ce

in
li
te
ra
cy

—
—

1
.1
2
2
4
*
**

0
.1
9
8
2

0
.2
3
6
9
*
*
*

0
.0
4
2

0
.5
9
7
9
*
*
*

0
.0
8
0
7

A
g
e

#
0
.0
9
1
1
*
*
*

0
.0
1
0
2

#
0
.0
9
1
8
*
**

0
.0
1
0
2

#
0
.0
8
8
3
*
*
*

0
.0
1
0
2

#
0
.0
9
1
6
*
*
*

0
.0
1
0
2

M
al
e

0
.3
8
6
8

0
.1
9
9
3

0
.3
6
6
4

0
.1
9
3
5

0
.4
9
7
5
*

0
.1
9
3
6

0
.3
7
0
7

0
.1
9
7
9

M
ar
it
al
st
at
u
s
(r
ef
:
m
ar
ri
ed
)

S
in
g
le

#
0
.6
8
4
4
*
*

0
.2
5
2
7

#
0
.7
6
2
4
*
*

0
.2
4
8
2

#
0
.7
3
8
7
*
*

0
.2
4
7
1

#
0
.6
8
9
8
*
*

0
.2
5
2
5

S
ep
ar
at
ed
/d
iv
o
rc
e/
w
id
o
w

#
0
.4
5
9
2

0
.3
4
8
6

#
0
.3
4
8
1

0
.3
4
1
7

#
0
.3
6
2
6

0
.3
4
2
1

#
0
.4
8
1
6

0
.3
4
7
1

P
re
se
n
ce

o
f
d
ep
en
d
en
t
ch
il
d
(r
ef
:
N
o
)

0
.3
5
1
9

0
.2
1
7
1

0
.3
5
0
8

0
.2
1
0
1

0
.3
5
3
9

0
.2
0
9

0
.3
6
2
7

0
.2
1
6
7

R
ac
e/
et
h
ni
ci
ty

(r
ef
:
W
h
it
e)

B
la
ck

#
0
.0
2
7
3

0
.2
8
3
6

0
.0
4
2
7

0
.2
7
4
2

0
.0
8
5
5

0
.2
7
1
7

#
0
.0
1
6
1

0
.2
8
3
5

H
is
p
an
ic

#
0
.3
7
8
8

0
.3
4
5
5

#
0
.3
7
9
1

0
.3
3
6
1

#
0
.3
3
5
2

0
.3
3
2
2

#
0
.3
5
4
1

0
.3
4
3
2

A
si
an
/o
th
er
s

0
.6
0
9
9

0
.3
1
5
7

0
.4
5
8
7

0
.3
0
7

0
.6
5
7
5
*

0
.3
0
7
6

0
.5
9
6
8

0
.3
1
5
1

E
d
u
ca
ti
o
n
(r
ef
:
h
ig
h
sc
h
o
o
l
d
ip
lo
m
a
o
r
lo
w
er
)

S
o
m
e
co
ll
eg
e

0
.5
0
8
2

0
.2
9
0
5

0
.5
1
2
7

0
.2
8
1
3

0
.5
0
9
3

0
.2
8
0
6

0
.5
0
3
4

0
.2
9
0
6

A
ss
o
ci
at
e
d
eg
re
e

0
.5
3
3
1

0
.3
5
1
1

0
.5
0
6
2

0
.3
3
7
9

0
.5
0
3
6

0
.3
4
0
1

0
.5
4
2
1

0
.3
5
1

B
ac
h
el
o
r
d
eg
re
e

0
.0
9
5
7

0
.2
9
3

0
.1
6
7
7

0
.2
8
4
1

0
.1
9
4
3

0
.2
8
5
6

0
.0
7
4
3

0
.2
9
1

P
o
st
-b
ac
h
el
o
r
d
eg
re
e

0
.3
3
2
2

0
.3
5
8
6

0
.3
3
8
7

0
.3
4
8
9

0
.3
8
0
3

0
.3
5
1
1

0
.3
1
0
2

0
.3
5
6
8

E
m
p
lo
y
m
en
t
st
at
u
s
(r
ef
:
fu
ll
-t
im

e
em

p
lo
y
ee
)

S
el
f-
em

p
lo
y
ed

#
0
.0
0
6
4

0
.3
0
3
2

0
.0
0
6
9

0
.3
0
0
6

0
.0
2
6
6

0
.2
9
8
3

#
0
.0
1
7
3

0
.3
0
2
6

P
ar
t-
ti
m
e
w
o
rk
er

0
.3
1
1
1

0
.3
4
0
1

0
.0
9
9
3

0
.3
3
1
0

0
.0
9
3
4

0
.3
2
8
3

0
.3
0
2
9

0
.3
4
0
4

H
o
m
em

ak
er

#
1
.4
0
1
7
*

0
.6
2
3
4

#
1
.4
6
1
7
*

0
.5
9
9
1

#
1
.5
8
9
8
*
*

0
.5
9
5
7

#
1
.4
0
3
5
*

0
.6
2
4
8

S
tu
d
en
t

#
0
.8
5
8
1

0
.4
5
5
0

#
0
.9
1
4
7
*

0
.4
3
9
9

#
0
.9
3
4
4
*

0
.4
4
7
0

#
0
.8
5
4
5

0
.4
5
4
6

D
is
ab
le
d

0
.5
0
4
8

0
.8
0
7
0

0
.1
8
7
5

0
.7
8
3
9

#
0
.0
0
0
5

0
.8
0
3
1

0
.5
2
7
3

0
.8
0
4
3

U
n
em

p
lo
y
ed

#
1
.2
4
4
6

0
.7
3
4
2

#
1
.1
2
2
6

0
.7
1
7
2

#
1
.0
8
9
7

0
.7
0
8
4

#
1
.2
3
8
3

0
.7
3
3
2

R
et
ir
ed

#
0
.1
4
0
2

0
.4
5
5
0

#
0
.2
2
6
3

0
.4
5
2
6

#
0
.2
9
2
6

0
.4
5
4
0

#
0
.1
4
1
1

0
.4
5
4
3

In
co
m
e
(r
ef
:
le
ss

th
an

$
3
5
,0
0
0
)

$
3
5
,0
0
0
–
$
4
9
,9
9
9

#
0
.6
8
6
5

0
.3
6
7
6

#
0
.6
6
9
7

0
.3
6
2
9

#
0
.7
6
3
8
*

0
.3
5
7
9

#
0
.6
8
6
0

0
.3
6
7
7

$
5
0
,0
0
0
–
$
7
4
,9
9
9

#
0
.3
4
4
2

0
.2
9
3
2

#
0
.4
0
9
4

0
.2
8
5
6

#
0
.4
8
9
4

0
.2
8
4
5

#
0
.3
4
8
3

0
.2
9
3
2

$
7
5
,0
0
0
–
$
9
9
,9
9
9

#
0
.5
8
5
9

0
.3
3
6
9

#
0
.5
7
2
8

0
.3
2
9
2

#
0
.7
1
9
2
*

0
.3
2
4
6

#
0
.5
9
1
7

0
.3
3
7
1

(c
o
n
ti
n
u
ed

o
n
n
ex
t
p
a
g
e)

K. T. Kim et al. / Financial Services Review 31 (2023) 121–132 127



T
ab
le

2
(C
on
ti
n
u
ed
)

V
ar
ia
b
le
s

(A
)
B
as
el
in
e
m
o
d
el

(B
)
O
ve
rc
on
fi
de
nc
e

in
de
x
1

(h
ig
h
su
bj
ec
ti
ve
,

lo
w
ob
je
ct
iv
e)

(C
)
O
v
er
co
nfi

d
en
ce

in
d
ex

2
a

(d
is
cr
ep
an
cy
)

(D
)
O
v
er
co
n
fi
d
en
ce

in
d
ex

3b

(r
es
id
u
al
)

C
o
ef
f.

S
E

C
o
ef
f.

S
E

C
o
ef
f.

S
E

C
o
ef
f.

S
E

$
1
0
0
,0
0
0
–
$
1
4
9
,9
9
9

#
0
.4
3
0
7

0
.3
4
9
7

#
0
.5
9
1
2

0
.3
4
4
6

#
0
.6
1
4
7

0
.3
4
1
6

#
0
.4
4
7
3

0
.3
4
8
8

$
1
5
0
,0
0
0
o
r
m
o
re

#
1
.2
6
9
3
*
*

0
.4
8
8
4

#
1
.3
4
5
8
*
*

0
.4
8
5
9

#
1
.4
3
0
6
*
*

0
.4
8
5
9

#
1
.2
7
1
6
*
*

0
.4
8
7
7

In
v
es
tm

en
t
as
se
ts
(r
ef
:
le
ss

th
an

$
5
,0
0
0
)

$
5
,0
0
0
–
$
2
4
,9
9
9

#
0
.7
1
7
8
*

0
.3
1
5
4

#
0
.7
8
0
5
*

0
.3
0
6
9

#
0
.6
9
2
9
*

0
.3
0
5
4

#
0
.7
1
4
2
*

0
.3
1
5
5

$
2
5
,0
0
0
–
$
4
9
,9
9
9

#
0
.1
9
1
8

0
.3
4
1
7

#
0
.1
9
0
7

0
.3
3
7
1

#
0
.0
9
3
0

0
.3
2
9
7

#
0
.1
9
5
4

0
.3
4
2
5

$
5
0
,0
0
0
–
$
9
9
,9
9
9

#
0
.5
0
0
3

0
.3
2
5
1

#
0
.3
1
3
5

0
.3
1
6
1

#
0
.2
1
0
6

0
.3
1
2
7

#
0
.5
0
6
9

0
.3
2
5
0

$
1
0
0
,0
0
0
–
$
2
4
9
,9
9
9

#
0
.5
7
5
1

0
.3
6
3
0

#
0
.2
8
9
7

0
.3
5
3
6

#
0
.2
6
9
8

0
.3
5
3
6

#
0
.5
8
3
5

0
.3
6
2
4

$
2
5
0
,0
0
0
–
$
4
9
9
,9
9
9

0
.0
9
6
7

0
.3
7
6
9

0
.3
0
8
4

0
.3
6
7
7

0
.3
8
2
5

0
.3
6
8
3

0
.0
9
4
6

0
.3
7
6
0

$
5
0
0
,0
0
0
–
$
9
9
9
,9
9
9

#
1
.3
4
5
3
*

0
.6
4
9
7

#
1
.0
0
1
9

0
.6
3
2
2

#
0
.9
7
9
0

0
.6
3
5
5

#
1
.3
8
0
9
*

0
.6
4
8
8

$
1
,0
0
0
,0
0
0
o
r
m
o
re

#
0
.0
8
2
2

0
.4
9
4
1

0
.4
3
6
9

0
.4
6
5
8

0
.5
1
0
9

0
.4
6
9
8

#
0
.1
3
4
9

0
.4
8
7
7

C
o
n
st
an
t

#
1
.7
6
5
2
*

0
.8
0
0
8

0
.3
1
7

0
.6
7
0
5

0
.4
3
1
7

0
.6
6
7
2

0
.5
8
2
1

0
.6
8
1
6

R
eg
io
n
al
fi
x
ed

ef
fe
ct
(S
ta
te
o
f
re
si
d
en
ce
)

Y
es

Y
es

Y
es

Y
es

M
o
d
el
fi
t

C
o
n
co
rd
an
ce

ra
te

8
9
.0
%

8
7
.5
%

8
7
.7
%

8
9
.0
%

N
o
te
s.
W
ei
g
h
te
d
re
su
lt
s.

a
D
is
cr
ep
an
cy

b
et
w
ee
n
su
b
je
ct
iv
e
an
d
o
b
je
ct
iv
e
fi
n
an
ci
al
li
te
ra
cy
.

b
W
e
re
g
re
ss

su
b
je
ct
iv
e
fi
n
an
ci
al
li
te
ra
cy

o
n
th
e
o
b
je
ct
iv
e
li
te
ra
cy

an
d
ta
k
e
th
e
re
si
d
u
al
as

an
o
v
er
co
n
fi
d
en
ce

m
ea
su
re
.

S
ig
n
ifi
ca
n
ce

le
ve
l:
*p

<
.0
5
,
*
*p

<
.0
1
,
*
*
*p

<
.0
0
1
.

128 K. T. Kim et al. / Financial Services Review 31 (2023) 121–132



unit increase in subjective literacy increased the odds by 80.6%. In Models B, C, and D, we
found strong positive associations between overconfidence (as specified three different ways)
and cryptocurrency investment. Investors with overconfidence in investment literacy (Model
B, index 1) had 3.1 times the odds of investing in cryptocurrency as investors who were not
overconfident. Additionally, a one unit increase in overconfidence index 2 (Model C)
increased the odds of investing in cryptocurrency by 26.7%, and a one unit increase in over-
confidence index 3 (Model D) increased the odds of investing in cryptocurrency by 81.8%.

Among control variables, age was negatively associated with the likelihood of investing
in cryptocurrency. Single investors were less likely to invest in cryptocurrency than married
couples. Homemakers and students had lower likelihood of investing in cryptocurrency than
full-time employee. Investors with the lowest amount of income were more likely to invest
in cryptocurrency than those with the highest income ($150,000 or more). Investors with the
lowest amount of investment assets (less than $5,000) were more likely to invest in crypto-
currency than those with the investment assets of $5,000–$24,999.

5. Discussion and implications

While investment in cryptocurrency has been rapidly growing, the academic research on
cryptocurrency investors is still understudied. To fill this gap, this study investigated the
associations between investment literacy and cryptocurrency investment of US investors.
One of the notable findings is that objective investment literacy is negatively while subjec-
tive literacy and overconfidence are positively associated with the likelihood of investing in
cryptocurrency. Further, we found various socioeconomic characteristics related to crypto-
currency investment.

In line with previous research on risky assets and stock investment, we confirmed a salient
effect of investment literacy on cryptocurrency investment. Given the high volatility and
anomalies in the cryptocurrency market, and the short history of the market, such invest-
ments do not seem prudent, so our finding that overconfident individuals are more likely to
hold them is not surprising. However, given the trend of increasing and considerable trading
volume in a market, our finding provides some implications for financial practitioners and
educators. Many studies have documented low level of financial literacy (e.g., Atkinson &
Messy 2012; Lusardi & Mitchell 2017) and significant divergence between objective and
subjective financial literacy (e.g., Kim et al., 2020; Robb et al., 2015). Use of financial
advice or delegation of financial decisions could be possible alternatives to accompany
improving one’s investment literacy and assessing their financial literacy level realistically.
Our study did not test this issue empirically, but the use of financial advice could be a com-
plementary and supplementary source of financial and investment literacy of investors.

This study has some limitations to be noted. First, given the cross-sectional design of the
NFCS Investor Survey, this study did not address any causal inferences between investment
literacy and investing in cryptocurrency. Ideally, analyses that use longitudinal dataset or ex-
perimental study are needed to address the causal inference. Second, the timeframe of cryp-
tocurrency investment is not specified in the NFCS survey, for example, “Have you invested
. . .” Also, the survey question does not ask specific type of cryptocurrency available in a
marketplace. Despite these limitations, this study provides a source of understanding the
divergence in investment literacy and its association with cryptocurrency investment.

K. T. Kim et al. / Financial Services Review 31 (2023) 121–132 129



Appendix

Sample characteristics

Variables Percentage

Age, mean (SD) 50.8 (17.6)
Gender

Male 56.7
Female 43.3

Marital status (ref: married)
Married 58.4
Single 25.4
Separated/divorce/widow 16.2

Presence of dependent child 31.3
Race/ethnicity

White 76.9
Black 8.3
Hispanic 6.9
Asian/others 7.9

Education
High school diploma or lower 20.4
Some college 24.1
Associate degree 10.8
Bachelor degree 26.6
Post-bachelor degree 18.1

Employment status
Full-time employee 42.9
Self-employed 9.0
Part-time employee 7.7
Homemaker 3.8
Student 2.5
Disabled 1.3
Unemployed 1.9
Retired 30.9

Income
Less than $35,000 15.6
$35,000–$49,999 12.3
$50,000–$74,999 23.9
$75,000–$99,999 18.9
$100,000–$149,999 18.3
$150,000 or more 11.0

Investment assets
Less than $5,000 12.4
$5,000–$24,999 15.1
$25,000–$49,999 10.8
$50,000–$99,999 14.9
$100,000–$249,999 17.0
$250,000–$499,999 14.2
$500,000–$999,999 8.3
$1,000,000 or more 7.3

Note. Weighted results.

130 K. T. Kim et al. / Financial Services Review 31 (2023) 121–132



References

Abreu, M., & Mendes, V. (2012). Information, overconfidence and trading: Do the sources of information mat-
ter? Journal of Economic Psychology, 33, 868–881. https://doi.org/10.1016/j.joep.2012.04.003

Ante, L., Fiedler, I., Meduna, M. V., & Steinmetz, F. (2022). Individual cryptocurrency investors: Evidence from
a population survey. International Journal of Innovation and Technology Management, 19. https://doi.org/
10.1142/S0219877022500080

Arias-Oliva, M., Pelegrı́n-Borondo, J., & Matı́as-Clavero, G. (2019). Variables influencing cryptocurrency use:
A technology acceptance model in Spain. Frontiers in Psychology, 10, 475. https://doi.org/10.3389/
fpsyg.2019.00475

Atkinson, A., & Messy, F. (2012). Measuring financial literacy: Results of the OECD/International Network on
Financial Education (INFE) Pilot Study. OECD Working Papers on Finance, Insurance and Private
Pensions, No. 15. OECD Publishing. https://doi.org/10.1787/5k9csfs90fr4-en

Avital, M., Beck, R., King, J. L., Rossi, M., & Teigland, R. (2016). Jumping on the blockchain bandwagon:
Lessons of the past and outlook to the future. In Proceedings of the Thirty Seventh International Conference
on Information Systems, Dublin (ICIS), (pp. 1–6). http://aisel.aisnet.org/cgi/viewcontent.cgi?article=1320&
context=icis2016

Barber, B. M., & Odean, T. (2000). Trading is hazardous to your wealth: The common stock investment perform-
ance of individual investors. The Journal of Finance, 55, 773–806. https://doi.org/10.1111/0022-1082.00226

Bharadwaj, S., & Deka, S. (2021). Behavioural intention towards investment in cryptocurrency: An integration
of Rogers’ diffusion of innovation theory and the Technology Acceptance Model. Forum Scientiae
Oeconomia, 9, 137–159.

Bouri, E., Gupta, R., & Roubaud, D. (2019). Herding behaviour in cryptocurrencies. Finance Research Letters,
29, 216–221. https://doi.org/10.1016/j.frl.2018.07.008

Chu, Z., Wang, Z., Xiao, J. J., & Zhang, W. (2017). Financial literacy, portfolio choice and financial well-being.
Social Indicators Research, 132, 799–820. https://doi.org/10.1007/s11205-016-1309-2

Chuen, D. L. K., Guo, L., & Wang, Y. (2017). Cryptocurrency: A new investment opportunity? The Journal of
Alternative Investments, 20, 16–40. https://doi.org/10.3905/jai.2018.20.3.016

Coin Market Cap. (2020, March 2). Cryptocurrency market capitalization. Available at https://coinmarketcap.
com/

da Gama Silva, P. V. J., Klotzle, M. C., Pinto, A. C. F., & Gomes, L. L. (2019). Herding behavior and contagion
in the cryptocurrency market. Journal of Behavioral and Experimental Finance, 22, 41–50. https://doi.org/
10.1016/j.jbef.2019.01.006

De Bondt, W. F., Muradoglu, Y. G., Shefrin, H., & Staikouras, S. K. (2008). Behavioral finance: Quo vadis?
Journal of Applied Finance, 18, 7–21.

Goetzmann, W. N., & Kumar, A. (2008). Equity portfolio diversification. Review of Finance, 12, 433–463.
https://doi.org/10.1093/rof/rfn005

Guiso, L., & Jappelli, T. (2008).Financial literacy and portfolio diversification. European University Institute
Working Paper ECO 2008/31. Available at: http://cadmus.eui.eu/handle/1814/9811.

Gupta, S., Gupta, S., Mathew, M., & Sama, H. R. (2021). Prioritizing intentions behind investment in cryptocur-
rency: A fuzzy analytical framework. Journal of Economic Studies, 48, 1442–1459. https://doi.org/10.1108/
JES-06-2020-0285

Hasso, T., Pelster, M., & Breitmayer, B. (2019). Who trades cryptocurrencies, how do they trade it, and how do
they perform? Evidence from brokerage accounts. Journal of Behavioral and Experimental Finance, 23, 64–
74. https://doi.org/10.1016/j.jbef.2019.04.009

Henry, C. S., Huynh, K. P., & Nicholls, G. (2018). Bitcoin awareness and usage in Canada. Journal of Digital
Banking, 2, 311–337.

Ibrahim, M. (2009). Theory of bounded rationality. Public Management, 91, 3–5.
Kim, K. T., Lee, J., & Hanna, S. D. (2020). The effects of financial literacy overconfidence on the mortgage de-

linquency of US households. Journal of Consumer Affairs, 54, 517–540. https://doi.org/10.1111/joca.12287

K. T. Kim et al. / Financial Services Review 31 (2023) 121–132 131



Kimball, M., & Shumway, T. (2006). Investor sophistication, and the participation, home bias, diversification,
and employer stock puzzles, Mimeo (Unpublished manuscript), Ann Arbor, MI: University of Michigan.

Liang, J., Li, L., Chen, W., & Zeng, D. (2019). Towards an understanding of cryptocurrency: A comparative
analysis of cryptocurrency, foreign exchange, and stock. In 2019 IEEE International Conference on
Intelligence and Security Informatics (ISI) (pp. 137–139). IEEE. https://doi.org/10.1109/ISI.2019.8823373

Lusardi, A., & Mitchell, O. S. (2007). Baby boomer retirement security: The role of planning, financial literacy, and
housing wealth. Journal of Monetary Economics, 54, 205–224. https://doi.org/10.1016/j.jmoneco.2006.12.001

Lusardi, A., & Mitchell, O. S. (2014). The economic importance of financial literacy: Theory and evidence.
Journal of Economic Literature, 52, 5–44. https://doi.org/10.1257/jel.52.1.5

Lusardi, A., & Mitchell, O. S. (2017). How ordinary consumers make complex economic decisions:
Financial literacy and retirement readiness. Quarterly Journal of Finance, 07, 1750008–1750031. https://
doi.org/10.1142/S2010139217500082

Nakamoto, S. (2008). Bitcoin: A peer-to-peer electronic cash system. Available at https://bitcoin.org/bitcoin.pdf
Robb, C. A., Babiarz, P., Woodyard, A., & Seay, M. C. (2015). Bounded rationality and use of alternative finan-

cial services. Journal of Consumer Affairs, 49, 407–435. https://doi.org/10.1111/joca.12071
Seay, M. C., Preece, G. L., & Le, V. C. (2017). Financial literacy and the use of interest-only mortgages. Journal

of Financial Counseling and Planning, 28, 168–180. https://doi.org/10.1891/1052-3073.28.2.168
Shin, S., Kim, K. T., & Seay, M. (2020). Source of information and selection of financial instruments. Journal of

Economic Psychology, 76, 102212. https://doi.org/10.1016/j.joep.2019.102212
Simon, H. A. (2000). Bounded rationality in social science: Today and tomorrow. Mind & Society, 1, 25–39.

https://doi.org/10.1007/BF02512227
Siriopoulos, C. (2021). Financial markets are not efficient: Financial literacy as an effective risk management tool.

International Journal of Business and Management Research, 9, 65–73. https://doi.org/10.37391/IJBMR.090110
Sorkin, A. R. (2021, March 10). It’s not just Bitcoin that’s huge. So is its carbon footprint. New York Times.
Tasca, P., Hayes, A., & Liu, S. (2018). The evolution of the bitcoin economy: Extracting and analyzing the net-

work of payment relationships. The Journal of Risk Finance, 19, 94–126. https://doi.org/10.1108/JRF-03-
2017-0059

Titov, V., Uandykova, M., Litvishko, O., Kalmykova, T., Prosekov, S., & Senjyu, T. (2021). Cryptocurrency
open innovation payment system: Comparative analysis of existing cryptocurrencies. Journal of Open
Innovation: Technology, Market, and Complexity, 7, 102. https://doi.org/10.3390/joitmc7010102

van Rooij, M., Lusardi, A., & Alessie, R. (2011). Financial literacy and stock market participation. Journal of
Financial Economics, 101, 449–472. https://doi.org/10.1016/j.jfineco.2011.03.006

Xia, T., Wang, Z., & Li, K. (2014). Financial literacy overconfidence and stock market participation. Social
Indicators Research, 119, 1233–1245. https://doi.org/10.1007/s11205-013-0555-9

Zhao, W. (2021, March 5). JPMorgan tells private wealth clients that bitcoin can be a portfolio diversifier ’if
sized correctly’. Available at https://www.theblockcrypto.com/post/97257/jp-morgan-bitcoin-deck-private-
client

Zhao, H., & Zhang, L. (2021). Financial literacy or investment experience: Which is more influential in cryptocur-
rency investment? International Journal of Bank Marketing, 39, 1208–1226. https://doi.org/10.1108/IJBM-11-
2020-0552

132 K. T. Kim et al. / Financial Services Review 31 (2023) 121–132



Assessing the relative need and demand for financial
education programs: A case study of graduate students

Brent J. Davisa,*, David P. Richardsona, Jason S. Seligmanb

aTIAA Institute, 8500 Andrew Carnegie Boulevard, Charlotte, NC 28262, USA
bInvestment Company Institute, 1401 H Street NW, Washington, DC 20005, USA

Abstract

We measure the relative need and demand for financial education programs among graduate stu-

dents at a large university system. We find that self-assessed and measured financial literacy is sig-

nificantly related to interest and demand for financial education. Individuals who self-report a high

level of financial literacy but have low measured financial literacy are significantly less likely to be

interested in financial education, while the opposite is true for financially literate individuals self-

reporting a low level of financial literacy. Our study adds to research showing the importance of

both believed and actual financial literacy measures and has implications for financial education pro-

grams. © 2023 Academy of Financial Services. All rights reserved.

JEL classifications: D14; G53

Keywords: Financial education; Financial literacy; Overconfidence

1. Introduction

Over the last several decades, there has been a substantial increase in the investment,
health, and longevity risk burdens borne by individuals. This shift has increased focus on the
importance of household financial literacy in managing these risks and enhancing financial
well-being. As noted by Lusardi and Mitchell (2014), a lack of financial literacy can have
long-term impacts on households’ financial well-being. For example, not properly managing

*Corresponding author. Tel.: +1-704-988-6271.
E-mail address: Brent.Davis@tiaa.org (B. J. Davis)

1057-0810/23/$ – see front matter © 2023 Academy of Financial Services. All rights reserved.

Financial Services Review 31 (2023) 133–150



debt or planning for retirement can have substantial impacts on financial well-being over the
lifecycle.

While recent research has documented the link between low financial literacy and finan-
cial fragility, many adults continue to have poor understanding of basic personal finance
concepts.1 However, effectively engaging individuals in improving their financial literacy
has proven challenging. In this paper, we conduct a case study on the need and demand for
financial education programming by graduate students in a large university system.
Leveraging the differences between self-assessed and actual financial literacy levels, we find
that students most in need of additional financial education are least likely to take advantage
of programming.

Graduate students are good candidates to receive financial education programming. They
are among the most receptive consumers of education, command commensurately higher
lifetime earnings postgraduation, and may be more likely to have complex financial planning
needs over the lifecycle. Many also have prior workforce experience, which may expose
them to employer-sponsored benefits and retirement savings plans. However, this prior
workforce participation may make graduate students more anxious about their financial se-
curity due to low earnings and savings and likely increased debt, during graduate studies.
Providing financial education programming during this time can help students gain confi-
dence in their long-term financial well-being.

Our interest in studying financial education programming is whether it is economically
rational for individuals to invest in financial education, and relatedly, can financial education
programs improve outcomes. Regarding the latter, evidence increasingly suggests that finan-
cial education affects outcomes across a variety of contexts.2 With respect to financial edu-
cation investment, Lusardi, Michaud, and Mitchell (2017) discuss that investment in
financial literacy should be a function of the associated expected benefits. They find that
low-education low-lifetime income groups might rationally choose not to invest in obtaining
financial literacy skills because of their larger relative reliance on public social welfare pro-
grams. Because our target sample is of high education and higher-than-median expected life-
time income, they should rationally choose to invest in financial literacy education.

We measure self-assessed and tested financial literacy for graduate students in a large
public university system. Similar to past studies, we find measurable weakness in the popu-
lations’ understanding of basic personal finance concepts, despite most students expressing
concern regarding their own personal finances. Regarding financial education engagement,
we document two main findings. The first is that interest in financial education relates signif-
icantly to an individual’s self-assessed level of financial knowledge relative to their actual
measured financial literacy level. Individuals who estimate they have a high level of finan-
cial literacy but perform poorly on a financial literacy quiz are significantly less likely to be
interested in financial education, while the opposite is true for financially literate students
who self-report a low level of financial acumen. We also find that while receptivity to finan-
cial education programming increases significantly with financial literacy levels, overall
engagement in financial education is low, despite nearly half of survey respondents signaling
interest in programming. The students most likely to indicate interest in financial education
are the ones with higher financial literacy and lower overconfidence in their financial liter-
acy. Our work adds to the small body of existing research, namely Allgood and Walstad

134 B. J. Davis et al. / Financial Services Review 31 (2023) 133–150



(2016) and Anderson et al. (2017), that perceived financial literacy and actual financial liter-
acy are dually important. Our results indicate that employers using financial education pro-
gramming need to adopt innovative engagement strategies to improve the financial literacy
and well-being of those employees who need it most.

2. Research design and methodology

Our educational protocol was developed using an early career workplace education offering
from a large financial services organization.3 We targeted the financial education offering to
the entire pool of graduate students with the following research protocol. At the beginning of
the academic year, and ahead of inviting the graduate student population to take our initial sur-
vey, we developed a series of prompts. These consisted of 5 ! 7 inch postcards announcing
the project with lighthearted financial literacy questions on a front side and the correct answer
along with a description of what to expect next on the back.4 We developed three distinct
cards, with the goal that students might compare the particular card they received to others
and discuss them. Regarding signals of validity, integrity, and quality of the effort, the cards
prominently featured the collaborating university and were placed in orientation packets for
incoming students by the university. The university also placed cards at each department’s
reception desk. A research assistant (RA) was assigned to attend graduate student council
meetings and other graduate student group meetings of various types across the university sys-
tem. In these student meetings, the RA was granted five-minute slots to discuss the project and
the potential benefits of participation and left cards at each meeting.5

Two weeks after the cards were distributed initial invitations for the online survey were
sent to all graduate students (17,819) in a large public university system. These invitations
were designed to resonate with the information on the cards we had just distributed. The sur-
vey was designed to record information on students’ individual and educational characteristics,
financial aspirations, personal financial concerns, self-assessed financial acumen, and a finan-
cial literacy quiz.6 Once the survey was completed, we provided quiz scores to students and
offered them the correct answers to missed questions. We then asked whether a respondent
was interested in taking part in a financial education seminar or webinar in the near future.

Our survey was open for one month between mid-September and mid-October, ahead of
midterm examinations. Over this four-week period the student received an initial invitation
and up to four reminders targeted to students who had not taken the survey nor opted out of
email engagement. Initial and reminder response survey engagement rates are shown in Fig. 1.

Our email prompts to engage the survey were successful with approximately 60% of the
survey sample engaged the survey following a reminder as seen with the spikes in the response
rate, supporting Dechausay et al.’s (2015) result that reminders can improve engagement.

From the initial invite population, 2,487 students (14%) engaged the survey. To set up the
second stage of our study we invited a matched, random subsample of our surveyed students
to a financial education seminar or webinar, whichever they preferred. We randomly
selected 1,632 students to match to invited and noninvited groups. Of the 1,632 students eli-
gible for invitation, roughly two-thirds (1,101) were invited to participate.7 Our invited

B. J. Davis et al. / Financial Services Review 31 (2023) 133–150 135



student group was designed to be a balanced representation of participants across several
dimensions: gender, degree of concern regarding financial matters, score on the financial lit-
eracy quiz, and whether or not the student was in a quantitative field.8 Invitations were also
balanced across those who did or did not initially indicate interest in the program. This
allowed students to change their mind if they later decided they wanted to attend. The invita-
tion email contained a link for those wishing to sign up and clicking the link brought the stu-
dent to a standard web-based submission form. In the invitation to participate, we offered
two mid-day and two early evening times for either a seminar or webinar at each campus in
the university system. We also offered lunch or dinner to seminar participants; something
our pilot run in the previous year had revealed as being important.

3. Results

In this section we present results from our survey and financial education engagement
protocol. We begin by examining the demographic, educational field, and financial literacy
characteristics of students who engaged the survey. Section 3.2 examines how financial liter-
acy (both self-assessed and actual) relate to interest in attending a financial education semi-
nar. In Section 3.3, we use regression analyses to examine correlates of interest in financial
education interest and engagement.

3.1. Survey sample characteristics

Table 1 presents characteristics of survey respondents. The average student who partici-
pated in the survey is nearly 30 years old and has almost 10 graduate course credits. Women
(57%) were more likely to participate compared to men, approximately a quarter are

Fig. 1. Survey response rates.

136 B. J. Davis et al. / Financial Services Review 31 (2023) 133–150



international students, 57% of the graduate students attend the flagship campus, and nearly
one in four graduate students work as research or teaching assistants. Half of the students in
the sample are pursuing a Master’s degree, over a quarter a doctorate, and other degree types
represent less than 10% of the survey respondents. About two-thirds of survey participants
have prior work experience.

We categorize a student’s program or major into four mutually exclusive groups: Liberal

Arts (e.g., liberal arts, humanities, language, music, and social sciences), STEM (e.g., sci-

ence, engineering, medicine, mathematics, and technology), Professional (e.g., public health,

public administration, education, nursing, law, and other (pre)professional programs), and

Business (e.g., economics, finance, business, and accounting). Enrollment in a professional

program represents the plurality of students surveyed (37%) followed by enrollment in

STEM (27%), Liberal Arts (20%), and Business (14%) programs.
Table 2 displays financial education engagement numbers. Of the nearly 18,000 students we

sent a survey, about 14% (2,487) responded. Of the representative 1,101 students to whom we
sent invitations for a financial education session, 16% (176) accepted. Among the accepted
group, 36% (64) attended a session with 44 opting for an in-person seminar and 20 for an online
webinar. Acceptance rates for students who initially indicated interest (25%) were significantly
higher than those who did not signal interest (8%). Program attendance rates for those initially
indicating interest were also higher (39% vs. 29%) but the difference is not significant.

We next tabulate personal financial characteristics and engagement across program type
and survey variables. Table 3 Panel A displays measurements of financial literacy, personal
financial concerns, and education engagement for all surveyed students by program type.
Financial Literacy Quiz Score is the average number of questions students answered cor-
rectly on our financial quiz (out of 12).9 Financial IQ is the average of students self-assessed
rating of how high their financial knowledge or IQ is, ranging from 1 (very low) to (7 very
high). Relative FinIQ is an individual’s relative financial knowledge calculated as the

Table 1 Characteristics of targeted graduate student population

Student survey sample summary statistics

Mean/
proportion

Standard
deviation

Obs Proportion Obs

Student characteristics Degree type
Age 29.5 7.5 2,487 Masters 0.50 2,487
Credits taken 9.8 5.1 2,487 Certificate 0.02 2,487
Married 0.37 2,306 Law 0.07 2,487
Female 0.57 2,487 Doctorate 0.27 2,487
International student 0.26 2,487 Medical 0.08 2,487
Research/teaching asst 0.24 2,487 Post-doc 0.01 2,487
Flagship campus 0.57 2,487 Other/non-degree 0.05 2,487
Prior work experience 0.68 2,288

Education program type
Liberal arts 0.20 2,487
STEM 0.27 2,487
Professional 0.38 2,487
Business and economics 0.14 2,487

B. J. Davis et al. / Financial Services Review 31 (2023) 133–150 137



relative difference between performance on the financial literacy quiz and one’s self-
assessed financial IQ level and normalized on a scale of 0 to 1. Values below 0.5 represent
overconfidence and values above 0.5 represent underconfidence. For example, a value of 0
indicates complete overconfidence in the self-assessment compared to their actual measured
financial literacy knowledge and corresponds to a student answering 0 out of 12 financial lit-
eracy questions correctly and indicating a very high level (7) of Financial IQ on the self-
assessment, whereas a value of 1 indicates complete underconfidence. A value of 0.5 indi-
cates neither over- nor underconfidence.

Our personal finance concern metric is measured on an intensity scale from 1 (no con-
cern) to 5 (great concern) across five areas (career goals, current finances, future finances,
owning a home, and retirement), with a possible range from a low of 5 to a high of 25, which
we normalized from 0 to 1. For engagement metrics, we list the percentage of students who
responded to the survey, indicated that they were interested in financial education, accepting
an invitation for an educational session, and attended a seminar or webinar (conditional on
accepting an invitation).

Table 3 shows the average student got 65% of the financial literacy questions correct,
with business students scoring significantly higher than any of the other program groups.10

The mean self-reported financial IQ was 4.7 out of 7, again with business students indicating
a significantly higher level of self-reported financial knowledge compared to nonbusiness

Table 3 Financial literacy, financial concern, and financial education engagement

Characteristic Overall Program type

Liberal arts STEM Business Professional

Financial literacy and concern
Financial literacy quiz score (0–12) 7.81 7.57 7.66 8.71 7.71
Financial IQ, self-assessed (1–7) 4.67 4.58 4.42 5.28 4.73
Relative FinIQ (0–1) 0.52 0.52 0.53 0.51 0.51
Personal finance concern (0–1) 0.72 0.73 0.70 0.71 0.73

Education engagement
Responded to survey 14% 14% 14% 14% 14%
Interested in financial education 48 47 48 46 50
Accept invite j invited 16 17 15 13 17
Attend financial education conditional on
accepting invite

36 51 27 42 33

Note. Means or percents reported.

Table 2 Financial education engagement results

Engagement Number Number invited Proportion

Survey respondents 2,487 17,819 0.14
Eligible for invite 1,101 1,632 0.67
Interested in Fin Ed j eligible 511 1,101 0.46
Accepted j invited 176 1,101 0.16
Attended j accepted invite 64 176 0.36
Conditional on indicating interest

Accepted invite 383 511 0.25
Attended j accepting invite 50 128 0.39

138 B. J. Davis et al. / Financial Services Review 31 (2023) 133–150



students (5.28 vs. 4.60 for nonbusiness students). The average student was slightly under-
confident in their financial knowledge. There are limited differences in the relative measure
by program type; however, STEM students display significantly greater underconfidence
compared with the rest of the surveyed population. We find no significant differences by
program for students who responded to the survey or for those who indicated interested in
attending a financial education seminar or for acceptance rates. For seminar attendance,
however, we find a significantly greater proportion of Liberal Arts students attended a semi-
nar compared to students in other programs.

Table 4 shows asset and debt characteristics of the surveyed students. We hypothesize
that greater participation in financial services and the incidence of debt would be positively
correlated to signaling interest in financial education. And having life insurance or an invest-
ment account may signal greater interest in financial planning over the life cycle. We include
the incidence of students with a checking account, savings account, investment account, or a
life insurance policy. Nine of 10 students surveyed have a checking account, with three in
four having a savings account. The third row list the proportion of students with an invest-
ment account, which we define as having a brokerage account, an IRA, or an employer-
sponsored retirement savings plan. This is owned by a minority of students (38%), but varies
significantly by program type, ranging from 22% for STEM students to 50% for business
students.

We posit debt should be positively correlated with financial education interest. Graduate
students with student loan debt need to manage loan repayments in conjunction with other
consumption, savings, and investment goals and in context of their postgraduate career out-
look. Over half of the sample has some form of student loan debt, and this varies signifi-
cantly by a student’s major field. Professional students were significantly more likely to
have any debt, both debt from graduate and undergraduate studies and from credit cards.

Table 4 Other student characteristics, assets and debt

Characteristic Overall Program type

Liberal arts STEM Business Professional

Banking, assets, insurance
Checking account 90% 93% 88% 88% 90%
Savings account 74 78 71 72 75
Investment account 38 38 22 50 44
Life insurance 34 29 23 38 42

Student debt
Only undergraduate 9 12 9 5 8
Only graduate 19 16 22 26 17
Both undergrad and grad 28 30 22 15 36
Any student debt 56 57 53 46 61

Other debt
Credit card 31 34 23 26 37
Auto 20 16 15 23 24
Mortgage 20 20 11 24 25

Home ownership
Home owner 22 21 12 27 28
Plan to purchase home 46 39 55 50 43

B. J. Davis et al. / Financial Services Review 31 (2023) 133–150 139



Credit card debt may demonstrate greater need for financial education since it relates to
household balance sheets and not to human capital acquisition or indicates the reliance on
high-cost credit card debt to finance education expenses.

Our final category is home ownership. About one in five already own a home and roughly
50% of students are planning to purchase one in the next 10 years, which varies significantly
by a student’s major field. We later control for this in our regression analysis because pur-
chasing a home involves a substantial amount of financial planning, and we expect those
planning to purchase a home to have greater interest in financial education.

3.2. Interest in financial education

In this section, we explore the relationship between self-assessed and measured financial
literacy and graduate student initial interest in attending a financial education seminar.
Nearly half (48%) of the students who participated in the survey indicated they were inter-
ested in financial education. In Table 5, we find differences for indicated interest across
nearly all individual characteristics. While individuals who do not interact with financial
institutions may gain marginally greater benefit from financial education, we find that those
without checking and savings accounts, or life insurance are significantly less interested in
financial education. However, on the liability side of household balance sheets, students
with student loan debt or credit card debt were more likely to be interested in the educational
offerings.

Table 5 Student characteristics by financial education interest

Survey characteristics Interested Not interested Sig

Financial literacy and concern
Financial literacy quiz score 8.31 7.19 ***
Financial IQ (self-assessed) 4.59 4.82 ***
Relative FinIQ 0.55 0.48 ***
Personal finance concern 0.75 0.69 ***

Banking, assets, insurance
Checking account 98% 81% ***
Savings account 83 65 ***
Investment account 42 34 ***
Life insurance 37 31 ***

Student debt
Only undergraduate 11% 7% ***
Only graduate 21 18 *
Both undergrad and grad 32 24 ***
Any student debt 64 48 ***

Other debt
Credit card 37% 26% ***
Auto loan 23 17 ***
Mortgage 20 20

Home ownership
Own home 21% 23%
Plan to own home 55 39 ***

Note. Means or percents shown.
* and *** indicates differences are significant at the 10% and 1% levels, respectively.

140 B. J. Davis et al. / Financial Services Review 31 (2023) 133–150



Graduate students signaling interest in financial education have significantly higher finan-
cial literacy, with interested students averaging about 1.2 more correct answers compared to
non-interested students. Further, non-interested students are significantly more overconfident
in their financial knowledge. This is also shown by the distribution of Relative FinIQ shown
in Fig. 2, which displays kernel density estimates of Relative FinIQ by indicated interest in
financial education. The two distributions are significantly different (p < .01), with non-
interested students having a greater estimated density in the overconfidence range (values
below 0.5). This result is economically meaningful as the mean difference of 0.07 in the rel-
ative measure between interested and non-interested students equates to 1.68 fewer ques-
tions answered correctly (out of 12) on our financial quiz for the relatively more
overconfident student given a fixed self-assessed score. Alternatively, a 0.07 difference
results in a 0.84 greater self-assessed score, given a fixed quiz score, for the relatively more
overconfident student.

Overconfident students with low measured financial literacy would arguably benefit the
most from financial education. Unfortunately, Table 6 and Fig. 3 indicates that these stu-
dents were least interested in improving their financial knowledge. In Table 6, we examine
the percentage of students interested in financial education by their self-assessed financial
IQ rating and how well they did on the financial literacy quiz, grouping the latter into four
categories. The data suggest there is a strong negative correlation between overconfidence
and interest in financial education.11 This finding is highlighted by Fig. 3, which displays a
wireframe surface plot on the data points shown in Table 6. Financial literacy and self-
assessed financial IQ are shown on the x and y axes, and the percentage of students who
indicated they are interested in financial education is shown on the z (vertical) axis. Here the
relationship is clear: students who have high financial literacy but self-assess a low level of
financial IQ are the most interested in financial education (often over 70%). Comparatively,
students with low financial literacy but self-assess a high level of financial IQ are less inter-
ested (generally less than 30% of the time).

Fig. 2. Kernel density estimates of Relative FinIQ by indicated interest in financial education.

B. J. Davis et al. / Financial Services Review 31 (2023) 133–150 141



To understand this relationship over all possible values, Fig. 4 displays predicted proba-
bilities of financial education interest. The predicted probabilities are generated using a sim-
ple logit model estimating the likelihood that students indicated interest in financial
education regressed on a student’s FinIQ and Financial Literacy Quiz Score, treating the two
exogenous variables as categorical variables.12 The full profile of predicted probabilities is
shown in Appendix Table A.1. Fig. 4 replicates and smooths the relationship shown in
Fig. 3. Generally, the predicted probability that a student signals interest increases signifi-
cantly in one’s measured financial literacy but significantly decreases as a student’s self-
assessment increases. Students with a combination of high self-assessed financial IQ and
low actual financial literacy are predicted to be the least likely to signal interest in financial
education. For example, a student answering all questions correctly but self-assesses the
lowest level of financial knowledge is predicted to signal interest with a probability above
60%, which decreases to 45% for a student self-assessing the highest level of financial
knowledge. By contrast, a student answering no questions correctly on the financial literacy

Table 6 Financial education interest by measured and self-assessed financial literacy

Percent interested in financial education

Self-assessed Percent of financial quiz questions correct

Financial IQ 0–25% 26–50% 51–75% 76–100% Mean N

1 (very low) 50% 80% 50% 0% 55% 20
2 29% 65% 72% 82% 63% 80
3 26% 61% 68% 70% 62% 236
4 20% 49% 61% 58% 56% 646
5 22% 30% 64% 59% 57% 481
6 6% 30% 57% 55% 51% 484
7 (very high) 22% 0% 50% 45% 42% 153
Mean 20% 55% 62% 39% 48%
N 221 257 974 1,035 2,100

Fig. 3. Surface plot of financial education interest by measured and self-assessed financial literacy.

142 B. J. Davis et al. / Financial Services Review 31 (2023) 133–150



quiz but is most confident in their financial knowledge, is predicted to signal interest with a
likelihood of under 6%, increasing to only 11% for a student self-assessing the lowest level
of financial knowledge. Program sponsors need to be innovative when thinking about how to
engage the population of students who would most benefit from financial education.

3.3. Regression analysis of financial education interest and engagement

We begin our regression analysis with first considering who takes the survey. Table 7
uses ordinary Probit regression estimating whether a student engages in the survey or not
using student and program characteristics, displaying marginal effects on the coefficients
and standard errors in parenthesis. Because those who do not take the survey do not offer us
data on their financial literacy, we only leverage the university system’s administrative data.
Women and student workers are significantly more likely to engage the survey. While there
is no significant relationship for graduate students at the flagship campus, student workers at
the flagship campus are significantly less likely to engage the survey, indicated by the inter-
action term in Model 2. These students may be more time constrained with their studies and
work duties than their counterparts. We find no significant effect for international students.
Students who have taken more graduate credits are significantly less likely to take the sur-
vey, but the marginal effect is small and has significant attenuation. There were no signifi-
cant differences in a student’s major subject area. When examining degree type, law
students were significantly less likely and doctorate students only marginally less likely to
engage in the survey, highlighting the possibility of time constraints for students in terminal
degree programs.

Table 8 estimates the likelihood that a student indicates interest in financial education
using ordinary Probit specifications showing marginal effects and standard errors in paren-
thesis. We display five specifications: Model 1 uses administrative data plus financial liter-
acy characteristics, and Models 2 and 3 add asset and debt characteristics. Model 4 adds

Fig. 4. Predicted probabilities of education interest by measured and self-assessed financial literacy.

B. J. Davis et al. / Financial Services Review 31 (2023) 133–150 143



T
ab
le

7
R
eg
re
ss
io
n
es
ti
m
at
es

o
f
su
rv
ey

en
g
ag
em

en
t

(1
)

(2
)

(3
)

(4
)

M
ar
.
co
ef
f.

S
ta
n
d
ar
d
er
ro
r

M
ar
.
co
ef
f.

S
ta
n
da
rd

er
ro
r

M
ar
.
co
ef
f.

S
ta
n
d
ar
d
er
ro
r

M
ar
.
co
ef
f.

S
ta
n
d
ar
d
er
ro
r

S
tu
de
n
t
ch
ar
ac
te
ri
st
ic
s

W
o
m
an

0
.0
3
0*
*
*

(0
.0
0
5
)

0
.0
3
0
*
*
*

(0
.0
0
5
)

0
.0
3
0
*
*
*

(0
.0
0
5)

0
.0
3
0
*
*
*

(0
.0
0
5
)

A
g
e

"
0
.0
0
0

(0
.0
0
0
)

0
.0
0
0

(0
.0
0
0
)

"
0
.0
0
0

(0
.0
0
0)

0
.0
0
0

(0
.0
0
0
)

In
tl
st
u
d
en
t

0
.0
0
9

(0
.0
0
6
)

0
.0
0
7

(0
.0
0
6
)

0
.0
0
8

(0
.0
0
6)

0
.0
0
6

(0
.0
0
6
)

M
ai
n
ca
m
p
u
s

0
.0
0
2

(0
.0
0
6
)

0
.0
0
9

(0
.0
0
6
)

0
.0
0
7

(0
.0
0
6)

0
.0
0
8

(0
.0
0
6
)

S
tu
d
en
t
te
ac
h
in
g
/r
es
ea
rc
h
as
st

0
.0
3
7*
*
*

(0
.0
0
7
)

0
.0
7
5
*
*
*

(0
.0
1
4
)

0
.0
7
7
*
*
*

(0
.0
1
4)

0
.0
7
3
*
*
*

(0
.0
1
4
)

E
d
u
c
cr
ed
it
s

"
0
.0
0
6*
*
*

(0
.0
0
1
)

"
0
.0
0
6
*
*
*

(0
.0
0
1
)

"
0
.0
0
6
*
*
*

(0
.0
0
1)

"
0
.0
0
5
*
*
*

(0
.0
0
2
)

ln
(e
d
u
c
cr
ed
it
s)

0
.0
6
1*
*
*

(0
.0
1
1
)

0
.0
6
1
*
*
*

(0
.0
1
1
)

0
.0
6
0
*
*
*

(0
.0
1
2)

0
.0
5
6
*
*
*

(0
.0
1
2
)

M
ai
n
C
am

p
u
s
*
S
tu
d
en
t
A
ss
t

"
0
.0
5
0
*
*
*

(0
.0
1
6
)

"
0
.0
4
9
*
*
*

(0
.0
1
6)

"
0
.0
4
6
*
*
*

(0
.0
1
6
)

P
ro
g
ra
m

ty
p
e
(b
as
el
in
e
¼

L
ib
er
al

A
rt
s)

S
T
E
M

"
0
.0
0
2

(0
.0
0
9)

P
ro
fe
ss
io
n
al

0
.0
0
3

(0
.0
0
8)

B
u
si
n
es
s

0
.0
0
7

(0
.0
1
0)

P
ro
g
ra
m

ty
p
e
(b
as
el
in
e
¼

M
as
te
rs
)

D
o
ct
o
ra
te

"
0
.0
1
2
*

(0
.0
0
7
)

L
aw

"
0
.0
3
3
*
*
*

(0
.0
0
9
)

O
th
er
/n
o
n
-d
eg
re
e

"
0
.0
0
0

(0
.0
1
3
)

C
er
ti
fi
ca
te

"
0
.0
2
7

(0
.0
1
8
)

P
o
st
-d
o
c

"
0
.0
1
7

(0
.0
3
2
)

M
ed
ic
al

"
0
.0
1
6

(0
.0
1
1
)

O
b
se
rv
at
io
n
s

1
7
,7
5
0

1
7
,7
5
0

1
7
,7
5
0

1
7
,7
50

N
o
te
.
P
ro
b
it
sp
ec
ifi
ca
ti
o
n
s
w
it
h
m
ar
g
in
al
co
ef
fi
ci
en
ts
re
p
o
rt
s
an
d
st
an
d
ar
d
er
ro
rs
in

p
ar
en
th
es
is
.

*
,
*
*
,
*
*
*
re
p
re
se
nt
s
si
g
n
ifi
ca
n
ce

at
th
e
1
0
%
,
5
%
,
an
d
1
%

le
v
el
s,
re
sp
ec
ti
v
el
y
.

144 B. J. Davis et al. / Financial Services Review 31 (2023) 133–150



T
ab
le

8
R
eg
re
ss
io
n
es
ti
m
at
es

o
f
se
m
in
ar

in
te
re
st (1
)

(2
)

(3
)

(4
)

(5
)

S
tu
de
n
t
an
d
p
ro
gr
am

ch
ar
ac
te
ri
st
ic
s

W
o
m
an

0
.0
2
3
(0
.0
2
2)

0
.0
2
1
(0
.0
2
3
)

0
.0
2
1
(0
.0
2
3)

0
.0
1
9
(0
.0
2
2
)

0
.0
0
1
(0
.0
2
2)

A
g
e

"
0
.0
0
4
(0
.0
0
2)
*
*
*

"
0
.0
0
4
(0
.0
0
2
)*
*

"
0
.0
0
3
(0
.0
0
2)
*
*

"
0
.0
0
1
(0
.0
0
2
)

"
0
.0
0
1
(0
.0
0
2)

In
t’
l
st
u
d
en
t

0
.0
3
5
(0
.0
2
7)

0
.0
2
9
(0
.0
2
8
)

0
.0
6
5
(0
.0
3
1)
*
*

0
.0
5
9
(0
.0
3
1
)*

0
.0
4
2
(0
.0
3
1)

M
ai
n
ca
m
p
u
s

0
.0
0
8
(0
.0
2
4)

0
.0
0
3
(0
.0
2
4
)

0
.0
0
9
(0
.0
2
4)

0
.0
0
5
(0
.0
2
4
)

0
.0
0
4
(0
.0
2
5)

R
es
ea
rc
h
/t
ea
ch
in
g
as
st

"
0
.0
0
1
(0
.0
2
6)

"
0
.0
0
8
(0
.0
2
6
)

"
0
.0
1
0
(0
.0
2
7)

"
0
.0
1
5
(0
.0
2
7
)

"
0
.0
1
7
(0
.0
2
7)

C
re
d
it
s

"
0
.0
0
5
(0
.0
0
2)
*
*

"
0
.0
0
6
(0
.0
0
3
)*
*

"
0
.0
0
6
(0
.0
0
3)
*
*

"
0
.0
0
6
(0
.0
0
3
)*
*

"
0
.0
0
6
(0
.0
0
3)
*
*

S
T
E
M

0
.0
4
4
(0
.0
3
3)

0
.0
4
3
(0
.0
3
3
)

0
.0
4
7
(0
.0
3
3)

0
.0
3
8
(0
.0
3
3
)

0
.0
3
5
(0
.0
3
4)

B
u
si
n
es
s

0
.0
1
4
(0
.0
3
8)

0
.0
2
1
(0
.0
3
9
)

0
.0
2
3
(0
.0
3
9)

0
.0
1
8
(0
.0
3
9
)

0
.0
3
6
(0
.0
3
9)

P
ro
fe
ss
io
n
al

0
.0
5
3
(0
.0
2
9)
*

0
.0
5
6
(0
.0
2
9
)*

0
.0
5
5
(0
.0
2
9)
*

0
.0
5
1
(0
.0
2
9
)*

0
.0
5
4
(0
.0
2
9)
*

F
in
an
ci
al
li
te
ra
cy

ch
ar
ac
te
ri
st
ic
s

F
in
an
ci
al
li
te
ra
cy

0
.0
4
3
(0
.0
0
4)
*
*
*

0
.0
4
3
(0
.0
0
4
)*
*
*

0
.0
4
4
(0
.0
0
4)
*
*
*

0
.0
4
4
(0
.0
0
4
)*
*
*

S
el
f-
as
se
ss
ed

fi
n
an
ci
al
IQ

"
0
.0
5
2
(0
.0
0
8)
*
*
*

"
0
.0
5
1
(0
.0
0
9
)*
*
*

"
0
.0
5
0
(0
.0
0
9)
*
*
*

"
0
.0
5
0
(0
.0
0
9
)*
*
*

R
el
at
iv
e
F
in
IQ

0
.8
5
3
(0
.0
7
2)
*
*
*

P
er
so
n
al
fi
na
n
ce

co
n
ce
rn

0
.4
5
8
(0
.0
5
5)
*
*
*

0
.4
6
1
(0
.0
5
5
)*
*
*

0
.4
5
2
(0
.0
5
5)
*
*
*

0
.4
3
1
(0
.0
5
6
)*
*
*

0
.4
2
6
(0
.0
5
6)
*
*
*

B
an
k
in
g,

as
se
ts
,
in
su
ra
n
ce

In
v
es
tm

en
t
o
r
re
ti
re
m
en
t

"
0
.0
1
9
(0
.0
2
6
)

"
0
.0
1
5
(0
.0
2
6)

"
0
.0
0
5
(0
.0
2
7
)

0
.0
1
0
(0
.0
2
6)

S
av
in
g
s
ac
co
u
n
t

0
.0
2
5
(0
.0
2
9
)

0
.0
2
8
(0
.0
2
9)

0
.0
2
8
(0
.0
2
8
)

0
.0
3
6
(0
.0
2
8)

L
if
e
in
su
ra
n
ce

"
0
.0
2
7
(0
.0
2
5
)

"
0
.0
3
1
(0
.0
2
5)

"
0
.0
1
9
(0
.0
2
5
)

"
0
.0
1
8
(0
.0
2
5)

D
eb
t

S
tu
d
en
t
d
eb
t
(u
n
d
er
g
ra
d
o
n
ly
)

0
.1
0
8
(0
.0
3
8)
*
*
*

0
.1
0
1
(0
.0
3
8
)*
*
*

0
.0
9
9
(0
.0
3
8)
*
*
*

S
tu
d
en
t
d
eb
t
(g
ra
d
o
n
ly
)

0
.0
2
7
(0
.0
3
0)

0
.0
1
9
(0
.0
3
0
)

0
.0
1
6
(0
.0
3
0)

B
o
th

u
n
d
er
g
ra
d
an
d
g
ra
d
d
eb
t

0
.0
6
0
(0
.0
2
9)
*
*

0
.0
5
1
(0
.0
2
9
)*

0
.0
4
3
(0
.0
2
9)

C
re
d
it
C
ar
d

"
0
.0
5
1
(0
.0
6
3)

"
0
.0
5
9
(0
.0
6
3
)

"
0
.0
5
0
(0
.0
6
3)

A
u
to

0
.0
3
2
(0
.0
2
7)

0
.0
4
2
(0
.0
2
7
)

0
.0
4
1
(0
.0
2
7)

H
o
m
e
o
w
n
er
sh
ip

O
w
n
h
o
m
e

"
0
.0
6
8
(0
.0
3
7
)*

"
0
.0
5
7
(0
.0
3
7)

P
la
n
to

p
u
rc
h
as
e
h
o
m
e

0
.0
5
3
(0
.0
2
7
)*
*

0
.0
6
1
(0
.0
2
7)
*
*

O
b
se
rv
at
io
n
s

2
,0
9
8

2
,0
9
8

2
,0
9
8

2
,0
9
8

2
,0
9
8

N
o
te
.
P
ro
b
it
w
it
h
m
ar
g
in
al
ef
fe
ct
s
sh
o
w
n
.

S
ta
n
d
ar
d
er
ro
rs
sh
o
w
n
in

p
ar
en
th
es
is
.
*
,
*
*
,
*
*
*
in
d
ic
at
es

si
g
n
ifi
ca
n
ce

at
th
e
1
0
%
,
5
%
,
o
r
1
%

le
v
el
.

B. J. Davis et al. / Financial Services Review 31 (2023) 133–150 145



home ownership characteristics, and Model 5 uses the relative FinIQ measure on the full
specification instead of quiz score and self-assessed financial IQ. Beginning with Model 1,
younger students are significantly more likely to signal interest, but this is not robust in later
specifications. Professional students are significantly more likely to signal interest; however,
this is only significant at the 10% level when controlling for debt characteristics. We find no
significant correlation for business students.13

In all specifications, financial literacy quiz scores and self-assessed Financial IQ have
large significant effects on the likelihood to signal interest. These effects pull in the opposite
directions, as discussed in Section 3.2., with predicted interest increasing in quiz score but
decreasing in the self-assessed measure. An additional question answered correctly on the
quiz score increases the estimated likelihood to indicate interest by 4%, while a one unit
increase in the self-assessed measure decreases the likelihood by 5%. The personal finance
degree of concern composite measure is significantly and positively related to indicating in-
terest, following our hypothesis. In Model 5, we use Relative FinIQ as a regressor, instead
of quiz score and the self-assess measure; and find estimated interest significantly increases
(decreases) as under- confidence (overconfidence) increases.

Adding individual asset and banking in Model 2 has no significant impact. Model 3
includes student loan debt as a categorical variable (with no student loan debt as the base-
line). Undergraduate student loan debt has a significant impact on the likelihood to be inter-
ested in financial education, compared to those without student loan debt. Although we do
not find a (robust) significant effect for either graduate debt only or both debt from under-
graduate and graduate school, the coefficient is positive—in the hypothesized direction.
Debt management is likely to become a larger concern for graduate students in the accumu-
lation phase of their lifecycle, especially since this group delays employment income, sav-
ings, and loan repayment before (re)entering the labor force, albeit at an expected relatively
higher salary. Credit card debt was not a significant correlate. This may be of concern
because some students may be using high-cost debt to finance part of their education and
this group would benefit from financial education. When including home ownership charac-
teristics in Model 4 we find those planning to purchase a home in the next ten years are sig-
nificantly more likely to signal interest, following our hypothesis.

4. Discussion

This paper examined financial education interest among graduate students in a large pub-
lic university system. We find a strong positive and significant correlation between under-
confidence (overconfidence) in self-measured financial knowledge and (lack of) interest in
financial education programming. This finding suggests the need for innovative engagement
strategies to identify and provide programming to individuals who would benefit the most
from improving their financial literacy.

The results speak to several components regarding the timing and delivery of financial
education. The first is whether the timing is optimal for graduate students to engage in
improving their financial literacy. Because many graduate students are close to (re)entering

146 B. J. Davis et al. / Financial Services Review 31 (2023) 133–150



the workforce, they may be focused on the near-term issues of graduating, finding a job, or
moving. This can make them subject to present bias through the belief that they have scant
time to devote additional resources to improving their long-term financial well-being.
Innovative engagement strategies, such a providing lunch or tchotchkes, may nudge active
participation in financial education. However, we find these nudges did not address to our
main finding that those confident in their financial knowledge but have low financial literacy
are significantly less interested in financial education.

How should these individuals be engaged in financial education? Mandatory financial edu-
cation could be one response, and many states have begun to institute mandatory financial lit-
eracy programs in high school. Stoddard and Urban (2020), Urban, Schmeiser, Collins, and
Brown (2018), and Collins (2013) find some benefits to mandatory education. However, there
needs to be further research in this area as there remain many open questions, including
whether the education benefits persist in later life, how such programs are implemented, what
is included in the content, and when in the lifecycle are they delivered, among others. With
most programs continuing to rely on voluntary education efforts, often offered by employer
benefit programs, designers and implementers of financial education will need to consider how
to attract individuals overconfident in their knowledge of personal finances.

Notes

1 For research on financial literacy see Yakoboski, Lusardi, and Hasler (2019), Clark,
Lusardi, and Mitchell (2017), Lusardi and Mitchell (2014), Lusardi, Mitchell, and
Curto (2014), Lusardi and Mitchell (2011), among others. For a broader discussion
on financial literacy, financial education, and economic outcomes see Hastings,
Madrian, and Skimmyhorn (2013).

2 Bernheim and Garrett (2003), Lusardi (2004), Maki (2004), and Bayer, Bernheim,
and Scholz (2009) have studied employer-sponsored financial literacy programs and
retirement preparedness. Other studies have found other positive benefits to financial
education programs (i.e., Clark et al., 2006; Skimmyhorn et al., 2016; and Seligman
and Bose 2012).

3 In the previous year, we piloted a similar survey and education to a small group of
students. While the test group gave good feedback and generally positive reviews,
we adjusted both our materials and engagement strategy to improve participation.

4 This postcard campaign is consistent with findings on the positive value of prompts
from Dechausay, Anzelone, and Reardon (2015).

5 The RA had previously served as President of the Graduate Student Council and
thus was a familiar and respected source of information across the body of groups
we engaged.

6 Lusardi, Mitchell, and Curto (2014), Schmeiser and Seligman (2013), and Knoll and
Houts (2012) have published work evaluating questions on measuring financial liter-
acy using three independent methodologies. We take our financial literacy questions
from this work, and consultation with financial counselors, and use a set of 12 ques-
tions from these studies.

B. J. Davis et al. / Financial Services Review 31 (2023) 133–150 147



7 A control group that was roughly one-third of eligible participants was not invited
so as to be able to carry forward with other research questions.

8 We defined a student as being in a quantitative field if the student’s program is in
economics, business, engineering, statistics, mathematics, physics, chemistry, or
computer science.

9 The financial literacy quiz covered questions on interest, inflation, and bond prices.
10 We control for degree type in our regression analysis.
11 Our relative measure cannot examine whether there are differences in student inter-

est across the entire cross product of the financial literacy score and self-assessed fi-
nancial IQ measure.

12 The overall model is significant (p < .01 with a x2 test). Moreover, the predicted
probabilities for each combination of FinIQ and Quiz score (91 combinations) are
all significant at the 5% level.

13 In separate regressions (not shown) we controlled for degree type and prior work ex-
perience; there were no significant differences.

Appendix

Table A1 Predicted probabilities of indicating interest in financial education by financial literacy and self-
assessed financial IQ

Financial literacy quiz score Self-assessed financial IQ

Very low 1 2 3 4 5 6 Very high 7

0 (0%) 11.1% 14.3% 13.7% 10.3% 10.5% 7.9% 5.9%
1 30.6 37.1 35.9 28.9 29.3 23.3 18.2
2 33.6 40.3 39.1 31.8 32.2 25.8 20.3
3 35.8 42.6 41.4 33.9 34.4 27.7 21.9
4 52.3 59.4 58.1 50.2 50.7 42.9 35.6
5 61.1 67.7 66.6 59.2 59.7 51.9 44.3
6 (50%) 54.1 61.2 60.0 52.1 52.6 44.8 37.3
7 65.9 72.1 71.1 64.1 64.5 57.1 49.4
8 63.0 69.4 68.3 61.1 61.5 53.9 46.2
9 64.5 70.8 69.8 62.7 63.1 55.6 47.9
10 62.0 68.6 67.5 60.1 60.6 52.9 45.2
11 62.2 68.7 67.6 60.2 60.7 53.0 45.4
12 (100%) 62.3 68.8 67.7 60.3 60.8 53.2 45.5

Note. Predicted probabilities from Logit model as described in Section 3.2, all are significant at the 5% level.

148 B. J. Davis et al. / Financial Services Review 31 (2023) 133–150



Acknowledgments

We thank Annamaria Lusardi and Carrie Houts for their comments on our financial literacy survey. We thank James Winbush,
Dan Rives, Keatrick Johnson, and others at Indiana University for their engagement and contributions. We have benefited from
the capable research assistance of Benjamin Bissette, Katherine Jones, and LaNita Rahjina. We acknowledge the financial
support of Indiana University in this work. Any errors are our own. The views are the authors’ and do not necessarily represent
the views of the TIAA, the TIAA Institute, or the Investment Company Institute.

References

Allgood, S., & Walstad, W. (2016). The effects of perceive and actual financial literacy on financial behaviors.
Economic Inquiry, 54, 675–697. Available at: https://doi.org/10.1111/ecin.12255

Anderson, A., Baker, F., & Robinson, D. (2017). Precautionary savings, retirement planning and misperceptions
of financial literacy. Journal of Financial Economics, 126, 383–398. Available at: https://doi.org/10.1016/j.
jfineco.2017.07.008

Bayer, P., Bernheim, D., & Scholz, J. (2009). The effects of financial education in the workplace: Evidence from a sur-
vey of employers. Economic Inquiry, 47, 605–624. Available at: https://doi.org/10.1111/j.1465-7295.2008.00156.x

Bernheim, D., & Garrett, D. M. (2003). The effects of financial education in the workplace: Evidence from a survey
of households. Journal of Public Economics, 87, 1487–1519. Available at: https://doi.org/10.1016/S0047-2727
(01)00184-0

Clark, R., Lusardi, A., & Mitchell, O. (2017). Employee financial literacy and retirement plan behavior: A case
study. Economic Inquiry, 55, 248–259. Available at: https://doi.org/10.1111/ecin.12389

Clark, R. L., d’Ambrosio, M. B., McDermed, A. A., & Sawant, K. (2006). Retirement plans and saving decisions:
The role of information and education. Journal of Pension Economics and Finance, 5, 45–67. Available at:
https://doi.org/10.1017/S1474747205002271

Collins, M. (2013). The impacts of mandatory financial education: Evidence from a randomized field study. Journal
of Economic Behavior & Organization, 95, 146–158. Available at: https://doi.org/10.1016/j.jebo.2012.08.011

Dechausay, N., Anzelone, C., & Reardon, L. (2015). The power of prompts: Using behavioral insights to encour-
age people to participate. OPRE Report 2015-75. Washington, DC: Office of Planning, Research and
Evaluation, Administration for Children and Families, U.S. Department of Health and Human Services.

Hastings, J., Madrian, B., & Skimmyhorn, W. (2013). Financial literacy, financial education, and economic outcomes.
Annual Review of Economics, 5, 347–373. Available at: https://doi.org/10.1146/annurev-economics-082312-125807

Knoll, M., & Houts, C. (2012). The financial knowledge scale: An application of item response theory to the
assessment of financial literacy. Journal of Consumer Affairs, 46, 381–410. Available at: https://doi.org/
10.1111/j.1745-6606.2012.01241.x

Lusardi, A. (2004). Saving and the effectiveness of financial education. In: O. Mitchell & S. Utkus (Eds.), Pension
design and structure: New lessons from behavioral finance (pp. 157–184). Oxford: Oxford University Press.

Lusardi, A., & Mitchell, O. (2011). The outlook for financial literacy. In: O. Mitchell & A. Lusardi (Eds.),
Financial literacy: Implications for retirement security and the financial marketplace (pp. 1–16). Oxford:
Oxford University Press.

Lusardi, A., & Mitchell, O. (2014). The economic importance of financial literacy: Theory and evidence. Journal
of Economic Literature, 52, 5–44. Available at: https://doi.org/10.1257/jel.52.1.5

Lusardi, A., Mitchell, O., & Curto, V. (2014). Financial literacy and financial sophistication in the older population.
Journal of Pension Economics & Finance, 13, 347–366. Available at: https://doi.org/10.1017/S1474747214000031

Lusardi, A., Michaud, P.-C., & Mitchell, O. (2017). Optimal financial knowledge and wealth inequality. The
Journal of Political Economy, 125, 431–477. Available at: https://doi.org/10.1086/690950

Maki, D. (2004). Financial education and private pensions. In W. Gale, J. Shoven, & M. Warshawsky (Eds.),
Private pensions and public policies (pp. 126–139). Washington, DC: Brookings Institution Press.

Schmeiser, M., & Seligman, J. (2013). Using the right yardstick: Assessing financial measures by way of finan-
cial well-being. Journal of Consumer Affairs, 47, 243–262. Available at: https://doi.org/10.1111/joca.12010

B. J. Davis et al. / Financial Services Review 31 (2023) 133–150 149



Seligman, J., & Bose, R. (2012). Learning by doing: Active employer sponsored retirement savings plan partici-
pation and household wealth accumulation. The Quarterly Review of Economics and Finance, 52, 162–172.
Available at: https://doi.org/10.1016/j.qref.2012.02.002

Skimmyhorn, W., Davies, E., Mun, D., & Mitchell, B. (2016). Assessing financial education methods: Principles
vs. rules-of-thumb approach. The Journal of Economic Education, 47, 193–210. Available at: https://doi.org/
10.1080/00220485.2016.1179145

Stoddard, C., & Urban, C. (2020). The effects of state mandated financial education on college financing behav-
iors. Journal of Money, Credit and Banking, 52, 747–776. Available at: https://doi.org/10.1111/jmcb.12624

Urban, C., Schmeiser, M., Collins, J., & Brown, A. (2018). The effects of high school personal financial educa-
tion policies on financial behaviors. Economics of Education Review, 78, 101786.

Yakoboski, P., Lusardi, A., & Hasler, A. (2019). Financial literacy in the United States and its link to financial
wellness. The 2019 TIAA institute-GFLEC personal finance index. Charlotte, NC: TIAA Institute.

150 B. J. Davis et al. / Financial Services Review 31 (2023) 133–150