Journal of The Japanese and International Economies: Tetsuya Kawamura, Tomoharu Mori, Taizo Motonishi, Kazuhito Ogawa

Journal of The Japanese and International Economies 60 (2021) 101131
Contents lists available at ScienceDirect
Journal of The Japanese and International Economies

journal homepage: www.elsevier.com/locate/jjie
Is Financial Literacy Dangerous? Financial Literacy, Behavioral Factors, and

Financial Choices of Households☆
Tetsuya Kawamura a, e, Tomoharu Mori b, e, Taizo Motonishi *, c, e, Kazuhito Ogawa d, e
a
Faculty of Economics and Business Management, Tezukayama University, Japan
b
College of Comprehensive Psychology, Ritsumeikan University, Japan
c
Faculty of Economics, Kansai University, Japan
d
Faculty of Sociology, Kansai University, Japan
e
Research Institute for Socionetwork Strategies, Kansai University, Japan
A R T I C L E I N F O A B S T R A C T
JEL classification: Using original purpose-built 2018 Japanese survey data, we estimate the financial behaviors and attitudes of
C83 households. We find that financial literacy plays an important and consistent role in financial decision-making.
D14 However, the actual behaviors are counter-intuitive: people with high levels of financial literacy tend to take too
G41
many risks, overborrow, and hold naive financial attitudes. That is, financial literacy tends to cause people to
Keywords: become daring and reckless toward some financial aspects. By contrast, financially literate people are better at
behavioral factor
retirement planning and are indifferent to gambling. Preferences such as risk and loss aversions and discount
consumer protection
financial education
factors, also play a role in financial choices.
financial literacy
household financial behavior
overconfidence
1. Introduction techniques and questionnaire methods to reveal individual preferences

and beliefs and have applied them to a wide range of behaviors (e.g.,
Owing to the growing need of people to save for their retirement, and Dohmen et al. (2011), Falk et al. (2018)). Since financial decisions can
the increased opportunities to undertake financial transactions as a be risky, ambiguous, and have a long-term; they can be strongly related
result of the development of financial technologies, an increasing to nonstandard as well as standard preferences.
number of households face complicated financial decisions. Researchers We investigate the effects of financial literacy and behavioral factors
have investigated the relationship between financial literacy and (i.e., preferences and beliefs) on households’ financial behaviors and
household decision-making to identify better financial education pol attitudes, while obtaining policy implications for financial education
icies for individuals. They find that financial literacy is related to more and consumer protection. Our analyses use a purpose-built original
investment in stocks (e.g., van Rooij et al. (2011)), higher returns (e.g., Internet survey of 5,848 respondents aged 20–80 years, conducted in
Bianchi (2018)), low-cost and moderate borrowing (e.g., Lusardi and March 2018. The survey questionnaire includes financial behaviors and
Tufano (2015)), and better retirement planning (e.g., Sekita (2011), attitudes, knowledge (e.g., financial literacy) and skills (e.g., calculation
Clark et al. (2012), van Rooij et al. (2012), and Anderson et al. (2017)). speed), standard preferences (e.g., risk aversion and discount rate),
Recent literature also focuses on economic preferences and beliefs. nonstandard preferences and beliefs (e.g., loss aversion and over
Economists and psychologists have been developing experimental confidence), and demographic factors.
☆
This work was supported by JSPS KAKENHI grant number 19K01769. We would like to thank Yusuke Kinari, Kazuhiko Nakahira, Kazuo Ueda, an anonymous
referee, and the seminar participants at the Japan Law and Economics Association meeting, the 17th International Conference of the Japan Economic Policy As
sociation, the MEW workshop, the International Symposium on Socionetwork Strategies in the Market of Data, the ISER-RISS Experimental Economics Workshop,
JEA Annual Meeting, and Waseda University for their comments and suggestions. We also thank Noriaki Kawamura of the Central Council for Financial Services
Information for giving us advice on the implementation of the financial literacy survey.
* Corresponding author. Tel.: Faculty of Economics, Kansai University, 3-3-35 Yamate-cho, Suita-shi, Osaka 564-8680.
E-mail address: tmoto@kansai-u.ac.jp (T. Motonishi).
https://doi.org/10.1016/j.jjie.2021.101131
Received 26 August 2020; Received in revised form 4 January 2021; Accepted 10 February 2021
Available online 13 February 2021
0889-1583/© 2021 Elsevier Inc. All rights reserved.
T. Kawamura et al. Journal of The Japanese and International Economies 60 (2021) 101131
A substantial amount of literature investigates the relationship be (2015) show that the same people tend to engage in high-cost
tween financial literacy, behavioral factors, and financial behavior.1 We borrowing. Sekita (2011) finds that those with high levels of financial
build upon previous literature in four ways. First, we evaluate financial literacy are more likely to have a retirement savings plan. Bianchi
behaviors and attitudes from several viewpoints: speculative invest (2018) argues that financially literate households earn higher financial
ment, share of risky assets, overborrowing, financial naivety, insuffi returns than illiterate ones. van Rooij et al. (2012) indicate that there is
cient retirement planning, and gambling interest. Second, we identify a “a strong and positive relationship between financial knowledge and
broad range of behavioral factors, such as the confidence gap in financial retirement planning.” The estimated effects of financial literacy on
literacy, present bias, loss aversion, S-shaped probability weighting, and financial behaviors in these studies generally favor households.
ambiguity aversion, as well as risk aversion and discount rate. These These survey-based studies are essential as they use data from re
nonstandard preferences and beliefs are essential in analyzing financial spondents from a wider variety of backgrounds compared to the
literacy because financial literacy education is usually considered a way experiment-based studies described below. The proof of causality in the
to prevent the irrational financial behaviors and attitudes of households. survey-based studies is, however, not definitive because they usually
Third, we use an original purpose-built survey to apply comprehensive cannot exploit exogenous variation in financial literacy.
and accurate measures of the determinants of financial behaviors and The second approach uses experimental or quasi-experimental
attitudes. Ten financial questions were formulated to measure financial techniques to evaluate the causal effect of financial education, finding
literacy, while incentivized questions were formulated to measure mixed results. Duflo and Saez (2003) find a significant effect of random
calculation speed and risk aversion. Fourth, the marginal effects of encouragement to attend a benefits information fair on the enrollment
financial literacy on financial behaviors is not assumed to be constant; rate of a tax-deferred account retirement plan. By evaluating
that is, we allow the effects of financial learning is different between randomly-assigned high school financial education program in Brazil,
novices and experts. Bruhn et al. (2016) concludes that while financial education improves
The regression results surprisingly show that high levels of financial savings and budgeting, it also leads to expansive credit use. Skimmy
literacy is associated with inappropriate financial behaviors and atti horn (2016)) shows, using a natural experiment, that the Personal
tudes, such as speculative investment, overborrowing, and financial Financial Management Course of the U.S. Army reduces some inappro
naivety. Thus, financial literacy tends to cause people to become daring priate credit card activities during the first year after the course, but not
and reckless in some financial aspects. In contrast, financially literate during the second year. Using state-level policy changes in the United
people are better at retirement planning and are indifferent to gambling. States, Brown et al. (2016), exploiting variation in financial education
The significant effects of the confidence gap in financial literacy reforms, argue that economical education leads to increased debt and
show that financial choices are determined by participants’ “perceived” repayment difficulties, whereas mathematical and financial education
and actual financial literacy. Moreover, a higher discount rate is asso have the opposite effects. The estimated effects of financial literacy on
ciated with overborrowing, insufficient retirement planning, and financial behavior in these studies are mixed.
increased gambling interest, whereas it leads to less financial naivety. Although the advantage of the second approach in identifying cau
Other behavioral factors generally tend to cause people to be less daring sality is apparent, this approach also has limits, for example that the data
and reckless financially: present bias, risk aversion, loss aversion, S- are obtained from people with a specific background, or people in a
shaped probability weighting, and ambiguity aversion. specific age group. Another limitation of these experimental studies is
Our results are especially noteworthy because the introduction of the degree of financial education intervention. Even a high school cur
financial education into secondary education is under consideration in riculum reform analyzed by Bruhn et al. (2016) could be limited
some countries, including Japan. It is true that young people are less compared to lifelong financial learning. In contrast, most existing
financially literate than their elderly counterparts. While financial lit survey-based studies analyze broad general population and evaluate the
eracy is widely believed to be effective in preventing households from effect of larger financial literacy differences. As we point out in Section
making inappropriate financial decisions, our results show the possi 3.3, people’s financial literacy improves almost linearly as they age from
bility of it leading to more inappropriate behaviors and attitudes, which their twenties to their seventies. Considering that financial behavior is
has important implications for financial education. relevant to the elderly, who generally have more assets than younger
We also show that the determinants of financial behavior are asso generations, it is crucial to evaluate the effects of large-scale lifelong
ciated with age and gender. The high level of financial knowledge of financial learning.
elderly males can be a contributing factor to financial fraud vulnera One of the limitations of the existing studies of both approaches is
bility and excessive risk-taking. Males tend to be more financially that they do not pay much attention to the stability of the gradient
literate, confident, and higher on discount rate than females on average. representing the relationship between financial literacy and behavior.
Younger generations are financially illiterate and less experienced while Existing survey-based studies assume a linear relationship. In experi
also less underconfident about financial literacy, and low on discount mental studies, it is usually not an easy task to evaluate different degrees
rate. of interventions. As Beshears et al. (2018) point out, a possible reason for
The remainder of this paper is organized as follows. Section 2 sum the mixed results in experimental studies is that what constitutes
marizes the literature related to this research. Section 3 explains the data “financial education” varies according to each study. One dimension of
and empirical methodology. Section 4 reports the estimation results, and the variability can include the degree of intervention. While Skimmy
Section 5 discusses the results and concludes. horn (2016), for example, evaluates the effect of an eight-hour financial
literacy course, Brown et al. (2016) examine the effect of high school
2. Related Literature curriculum reforms.
This study applies the first approach to evaluate the effects of lifelong
Existing research on the relationship between financial literacy and financial learning by using a purpose-built, original 2018 Japanese
household financial behaviors and attitudes use either survey-based or survey of the broad, general population. The survey includes ten
experimental approaches. Studies adopting the first approach show that financial literacy questions, which enables us to consider varying gra
financial literacy plays a vital role in the financial decisions of house dients of the relationship between financial literacy and behavior to
holds. van Rooij et al. (2011) conclude that people with low financial evaluate the effects of financial literacy. Moreover, although most
literacy are less likely to invest in stocks while Lusardi and Tufano existing studies analyze only a few types of financial behaviors and at
titudes, we investigate six types, including speculative investment, share
of risky assets, overborrowing, financial naivety, retirement planning,
1
See Beshears et al. (2018), for example. and gambling interest.
2
As for behavioral factors related to financial literacy, Anderson et al. asset holdings. Kadoya and Khan (2017) and Kadoya and Khan (2020)
(2017) focuses on actual and perceived literacy using data from Link investigated the factors of financial literacy such as demographic vari
edIn members and find that the effects of perceived literacy on pre ables and behavioral parameters. Kadoya et al. (2017), Sekita et al.
cautionary savings and retirement planning are greater than those of (2018), Yan et al. (2019), Sekita (2020), Fujiki (2020b), and Fujiki
actual literacy. Allgood and Walstad (2016), using data from National (2020c) examined the relationship between financial literacy and
Financial Capability Study conducted by FINRA, conclude that financial asset holdings. Kinari and Tsutsui (2009), Kitamura and
”perceived financial literacy may be as important as actual financial Nakajima (2010), Nogata and Takemura (2017), Ohtake and Akesaka
literacy.” Considering the fact that there are many existing studies on the (2017), Sekita et al. (2018), and Sekita (2020) examined the relationship
effects of financial literacy that do not measure perceived literacy, those among financial assets or the share of risky assets, behavioral biases, and
results are significant. preference parameters. Many studies show that overconfidence is posi
Many studies that use Japanese data have investigated the rela tively correlated with financial assets or the share of risky assets. In
tionship between financial literacy, individual preferences, and financial addition, many studies also show that low risk aversion is associated
with high asset value and the share of risky assets. For time preference,
the results vary from study to study and are not stable. The results of the
Table 1 share of risky assets in our study are generally consistent with these
Survey Questions for the Dependent Variables
results.
Categories Questions We build upon the previous literature on the effects of financial lit
Speculative SI1 You have experience dealing in the foreign eracy in the following four ways. First, we evaluate financial behaviors
Investment exchange (FX) market, trading stocks on and attitudes from several viewpoints: speculative investment, share of
margin, or trading futures or options. [Y] risky assets, overborrowing, financial naivety, insufficient retirement
SI2 You have experience trading cryptocurrencies
planning, and gambling interest. Second, we identify a broad range of
such as Bitcoin. [Y]
Over Borrowing OB1 You have used revolving credit through a behavioral factors. Third, we use an original purpose-built survey.
credit card. [Y] Fourth, the marginal effects of financial literacy on financial behaviors is
OB2 You have borrowed money through consumer not assumed to be constant.
financing. [Y]
Apart from the studies on the effects of financial literacy, the rest of
OB3 You have been denied a withdrawal from your
bank account because of insufficient funds. [Y] this section focuses on research on the effects of other behavioral factors
OB4 You have been refused a credit card or credit on financial behaviors. There is a substantial amount of this type of
card renewal. [Y] research, partly because the recent financial crisis has forced economists
OB5 You have been stuck paying down debt. [Y] and policymakers to pay attention to financial consumer protection.
OB6 You have borrowed money from a loan shark.
Shui and Ausubel (2005) find that consumers have a severe self-control
[Y]
Financial naivety FN1 You feel attracted to arrangements in which problem when they are faced with an introductory offer that presents a
trusted financial experts manage your finances lower interest rate within a shorter duration. There is a debate on the
for a 0.2% monthly fee. [Y] consequences of the effect of high-interest consumer credit such as
FN2 You feel attracted to life insurance and casualty
payday loans (e.g., Melzer (2011), Morse (2011)).
insurance policies that have no cash value at
maturity. [N]
Much financial economics and household finance research focuses on
FN3 You feel attracted to casualty insurance that risk preference as the central parameter in financial decisions. Donkers
offers compensation even for small damages. and van Soest (1999) and Dohmen et al. (2011) use subjective measures
[Y] of risk preference and relate them to financial behaviors. Guiso and
FN4 You feel attracted to investments described as
Paiella (2008) find that risk aversion, elicited by a direct question of
having a guaranteed principal and a high rate
of return. [Y] willingness-to-pay for security, is associated with income uncertainty
Insufficient RP1 You always scan the contents of letters about and liquidity constraints.
Retirement your public pension. [N] Time preferences are often mentioned in the context of debt
planning
borrowing. Using an incentivized experiment, Meier and Sprenger
RP2 You have checked your estimated pension
benefits. (For those already receiving such
(2010) find that present biases correlate with credit card debt, and
benefits, “you checked your benefits before Meier and Sprenger (2012) demonstrate that discount factors correlate
you began receiving them.“) [N] with creditworthiness and repayment decisions. In association with
RP3 You always know how many assets you have. financial literacy, Meier and Sprenger (2013) show that individuals who
[N]
have high discount factors are more likely to acquire personal financial
RP4 You always know how many liabilities you
have. [N] information.
RP5 You understand your yearly income and Ambiguity aversion is also highly mentioned because financial risks
expenditures (savings accounts, etc.). [N] often include some ambiguities. Dimmock et al. (2016) show that am
RP6 You generally know the value of your biguity aversion elicited by the Ellsberg urn experiment is negatively
estimated severance package. [N]
RP7 You have created your own financial plan for
correlated with stock market participation and the fraction of financial
your retirement. [N] assets in stocks. Numerous studies investigate the relationship between
RP8 You have asked an expert to create a financial overconfidence and financial behaviors (e.g., Biais et al. (2005), Deaves
plan for your retirement. [N] et al. (2009), Grinblatt and Keloharju (2009), Fellner-Röhling and
RP9 You have a financial plan that allows you not to
Krügel (2014), and Kinari (2016)). Most of them find results that are
worry about running out of funds, even if you
live 10 years longer than the average lifespan. consistent with economic theories.
[N] Recently, a growing number of studies in various fields have begun to
RP10 You are proactive about collecting information investigate the relationship between economic preferences, psycholog
on financial planning for your retirement from ical biases, and household behaviors using experiments or surveys. For
books, magazines, television, the Internet, etc.
[N]
example, Dohmen et al. (2011) and Sutter et al. (2013) investigate the
Gambling interest GI1 You have no interest in any kind of lottery relationship with a wide range of individual behaviors, including
(Jumbo Lottery, LOTO, Numbers, etc.). [N] financial ones. Barsky et al. (1997), Anderson and Mellor (2008), and
GI2 You have no interest in pachinko, horse racing, Goldzahl (2017) investigate the relationship in health economics, Burks
cycle racing, boat racing, or auto races. [N]
et al. (2009) and Fouarge et al. (2014) in labor economics, and Liu
3
(2013) in development economics. Table 2

There are methodological debates on the estimation of preferences Survey Questions for Some of the Independent Variables
and beliefs. Although revealed preferences elicited by incentivized ex Categories Questions
periments based on induced value theory (Smith (1976, 1982)) are
Financial literacy FL1 You have a normal savings account with
preferred in general, Mata et al. (2018) argues that stated preferences questions 10,000 yen and an interest rate of 2% per year.
are more stable than revealed one for risk preference. Horton et al. If you do nothing to change the balance, how
(2011) and Arechar et al. (2018) propose that the use of sophisticated much money will be in the account after one
Internet surveys or crowd-sourcing methods can mitigate physical and year? [1. 10200 yen] 2. I am not sure.
FL2 You have a normal savings account with
monetary costs of incentivized experiments. 10,000 yen and an interest rate of 2% per year.
If you do nothing to change the balance, how
3. Data and Methodology much money will be in the account after five
years? [1.More than 11,000 yen] 2. Exactly
11,000 yen. 3. Less than 11,000 yen. 4. I am not
3.1. Data
sure.
FL3 Your savings account has an interest rate of 1%
This study uses purpose-built survey data from the the Research per year, and the rate of inflation (the rate at
Institute for Socionetwork Strategies (RISS) of Kansai University. The which the cost of living increases) is 2% per
Institute conducts a series of web-based surveys on an irregular basis. year. After one year, which of the following
will you be able to buy with the money in the
The survey data were collected in March 2018 from a pool of 350,000 account? 1. More than you can now. 2. The
Japanese respondents, members of MyVoice, a web survey company in same as you can now. [3. Less than you can
Japan. The survey participants earn reward points consisting of ques now.] 4. I am not sure.
tionnaire participation fee of 90 points, and bonus points from the FL4 When interest rates increase, what happens to
bond prices? 1. They increase. 2. They do not
calculation and the risk aversion questions. Reward points can be
change. [3. They decrease.] 4. There is no
exchanged for universal gift certificates, book cards, PeX points, Amazon relationship between the interest rate and the
gift certificates, WebMoney, App Store money, or iTunes codes.2 cost of bonds. 5. I am not sure.
The company informs the registered members about the start of the FL5 When taking out a mortgage, the monthly
survey and recruits respondents according to the predetermined criteria payment is greater for a 15-year loan than for a
30-year loan, but there is less total interest paid
of 282 prefecture–age–gender groups (47 prefectures and age groups of during the repayment period. Is this statement
20s, 30s, 40s, 50s, 60s, and 70s and over). It should be noted that the correct or incorrect? [1. Correct] 2. Incorrect 3.
web survey sample is not immune to selection bias because respondents I am not sure.
are limited to the members of the web survey company, and there is a FL6 When investing, splitting an investment
between a number of companies is a safer way
self-selection process of the members to respond to a survey. To mini
to obtain returns than investing everything in a
mize the selection bias, we match the age, gender, and location distri single company. Is this statement correct or
bution of the respondents to that of the general population. The share of incorrect? [1. Correct] 2. Incorrect 3. I am not
each group is determined by the national population share of the pre sure.
fecture multiplied by the national population share of the age–gender FL7 When the value of the yen increases relative to
other currencies, what happens to the yen-
group. denominated value of financial assets held in
Bethlehem (2010) points out that access to the Internet and foreign currency? 1. The value increases. 2.
self-selection in the recruitment process can be sources of selection bias. The value does not change. [3. The value
Our sample’s average household income is 5.77 million yen, which is decreases.] 4. I am not sure.
FL8 When financial assets are sold quickly, they
close to the average household income of 5.52 million yen obtained from
tend to be sold at lower prices. Between real
13,000 random samples of entire households in the same year estate and government bonds, which
(Comprehensive Survey of Living Conditions by the Ministry of Health, experiences a greater decrease in price when
Labour and Welfare). This average income similarity implies that dif sold in this manner? [1. Real estate experiences
ferences in Internet access due to the income gap between our sample a greater decrease.] 2. The decrease is the
same. 3. Government bonds experiences a
and the general population are not significant.
greater decrease. 4. I am not sure.
The survey questions are tailored by the authors to capture financial FL9 There is no particular need to hold insurance
behaviors and attitudes, financial literacy, knowledge and skills, pref policies for events that have a very low chance
erences, beliefs, and other individual characteristics. The key survey of occurring. Is this statement correct or
incorrect? 1. Correct [2. Incorrect] 3. I am not
questions are presented in Tables 1 and 2.
sure.
FL10 Investments that are expected to have high
3.2. Empirical variables returns tend to also have considerable risks. Is
this statement correct or incorrect? [1. Correct]
2. Incorrect 3. I am not sure.
3.2.1. Financial behaviors and attitudes
Financial FE1 Experience in jobs related to financial matters,
Financial behaviors and attitudes are captured by the six dependent experience such as banks, brokerages, insurance, real
variables: speculative investment, share of risky assets, overborrowing, questions estate, etc. [Y]
financial naivety, insufficient retirement planning, and gambling interest. FE2 Experience handling finances for a company
These variables aim to capture potentially problematic financial be [Y]
Confidence in CFL FL1 to FL10 have dealt with financial literacy.
haviors and attitudes of the respondents. Financial behaviors and atti
financial literacy How many of these questions do you think you
tudes variables other than share of risky assets are created from answers were able to answer correctly? I was able to
to the questions presented in Table 1. The values of these variables are correctly answer (please select) questions.
the share of questions with the choices of answers in the square brackets Loss aversion LA There is a gambling game in which a coin is
question flipped and you receive a certain sum of money
in Table 1 for the corresponding group of questions.
if heads appear; however, you must pay 10,000
yen if tails appear. Assuming heads would
appear, how much money would you have to
2
The list of the gifts exchangeable for points is shown at http://www.myvoi (continued on next page)
ce.co.jp/voice/about/point.html
4
Table 2 (continued ) makes attaining competitive pricing complicated and difficult for the
Categories Questions insurance policy. Kunreuther et al. (2013) also argue that the popularity
of insurance policies with gaps in coverage is another example of
receive to participate in this game? Please
select the lower limit and assume that there are
anomalies in the insurance market. Some policies cover only minor
no legal problems with playing this game. damage but not catastrophic ones. Considering the costs incurred by the
Ambiguity aversion AA There are two urns, A and B, and each holds insurer and the effect of adverse selection, obtaining insurance for small
question 100 balls. The balls are either red or white. In damages is usually not reasonable. Question FN3 captures the re
Urn A, there are 50 red and 50 white balls, but
spondents’ preference for these unnecessary and possibly overpriced
the ratio of red to white balls in Urn B is
unknown. You must select one urn and draw insurance policies. Question FN4 test the respondents’ tendency to be
one ball out of that urn. If you draw a red ball, attracted to investment opportunities that seem unbelievable. Although
you will receive 10,000 yen. If you draw a an investment with a guaranteed principal and a high rate of return is
white ball, you will receive nothing. Which urn considered fraudulent advertising in financial markets, 11% of the re
do you choose?
spondents stated that they are attracted to these opportunities.
Insufficient retirement planning reflects the lack of preparation for
Speculative investment captures potentially risky financial behaviors aging. The questions used for creating the variable are RP1-RP10 in
by asking the respondents whether they have experience in trading Table 1. Insufficient retirement planning is the share of negative responses
Forex (FX), stocks on margin, futures, options, or Bitcoin. The questions to the questions. Under the universal public pension system, all Japanese
used for creating the variable are SI1 and SI2 in Table 1. Speculative individuals aged 20 and above receive annual statement about their past
investment is the share of affirmative responses to the questions. pension premium payments. Question RP1 in Table 1 asks whether the
Although these financial instruments are not necessarily speculative if respondents always review these statements. Question RP7 asks about
they are used prudently in appropriate occasions, for most households the planning behavior for retirement, with 13% of the respondents
they are still speculative as they are leveraged investments on volatile stating that they have created their own financial plan for retirement.
financial assets. FX is a form of foreign currency speculation in Japan Gambling interest measures attitudes against public lotteries and legal
that allows investors (often called Mrs. Watanabe) to trade at the gambling. The questions used for creating the variable are GI1 and GI2
maximum leverage of 25. Cryptocurrencies emerged as an alternative to in Table 1. Gambling interest is the share of negative responses to the
FX in recent years.3 questions. In Japan, gambling is forbidden except for state-controlled or
Share of risky assets is the share of individual stocks and foreign- decriminalized gaming. Question GI1 in Table 1 asks about interest in
currency-denominated assets in the house-hold financial assets. The public lotteries. Question GI2 refers to public gambling and pachinko, a
number is obtained from our survey question about asset shares. Japanese pinball game. For the purpose of giving an intuition about the
Overborrowing reflects high-interest or excessive borrowing. The penetration of gambling in Japan, figures on the size of gambling in each
questions used for creating the variable are OB1-OB6 in Table 1. Over category are given below. Annual sales of public lotteries, public
borrowing is the share of affirmative responses to the questions. In gambling, and pachinko in Japan were 0.16%, 1.0%, and 3.6% of GDP in
question OB1, “Revolving payment” (ribo-barai in Japanese) is a com 2018, respectively.
bination of revolving credit and a monthly minimum payment. The in
terest rate for a revolving payment by a credit card is approximately 3.2.2. Knowledge and skills
15%. Only 20% of the respondents stated that they had ever used Financial literacy, financial education at university, school education,
revolving payments.4 Another form of high interest rate borrowing is financial experience, and calculation ability are included in our financial
consumer financing in question OB2, for which the interest rate is about behavior regressions as knowledge and skills variables.
18%, equal to the legal ceiling on interest rates in Japan. Insufficient Financial literacy is the share of correct responses to the 10 financial
funds (question OB3), credit card renewal refusal (question OB4), questions FL1-FL10 presented in Table 2. The questions ask about in
default (question OB5), and borrowing from a loan shark (question OB6) terest calculation (FL1), compound interest calculation (FL2), inflation
are also included in overborrowing. (FL3), bond price (FL4), mortgage (FL5), diversified investment (FL6),
Financial naivety measures unsophisticated or ignorant attitudes of foreign exchange rate (FL7), liquidity (FL8), insurance (FL9), and risk-
households against potential rip-offs that are hyperbole or unnecessarily return tradeoff (FL10).
complicated. The questions used for creating the variable are FN1-FN4 Questions FL2, FL3, and FL6 correspond to the Big Three questions
in Table 1. Financial naivety is the share of naive responses to the by Lusardi and Mitchell (2011). In FL2, the amount of money is modified
questions. Question FN1 captures the indifference to financial fees. A to make the problem reasonably difficult, and it is presented in Japanese
0.2% monthly fee is equal to an annual fee of approximately a 2.4%, yen terms. FL3 is the same as the inflation question of the Big Three
which is as high as the average ROA of 2–3% for companies in the Nikkei questions. In FL6, we replaced “mutual fund” in the Big Three questions
225 during the 2010s. Question FN2 captures the respondents’ indif with its explanation, “splitting an investment between a number of
ference to the complexities of insurance policies. Kunreuther et al. companies” considering the low recognition of mutual funds in Japan.
(2013) present return of premium life insurance policies as an example Financial education at university is a dummy variable taking the value
of complex types of coverage, which is often regarded by financial ad of 1 if the subject graduated with a degree in economics, business, or
visors as questionable because the combination of insurance and savings commerce.
School education takes 0 for an junior high school diploma or less, 1
for a senior high school diploma, 2 for a junior college degree, 3 for a
college degree, 4 for a master’s degree, and 5 for a doctor’s degree.
3
A survey conducted by a cryptocurrency exchange in 2019 shows that only Financial experience is the share of affirmative responses to the
19% of cryptocurrency holders intend to use cryptocurrencies for making financial experience questions FE1–FE2 presented in Table 2.
payments (GMO Coin Inc, https://coin.z.com/jp/news/2019/03/1479/, in Calculation ability is measured by the logarithmic value of compu
Japanese).
4 tational speed of giving the correct answer to the five questions that
The use of the revolving payment in credit cards is less common in Japan: in
require choosing a row or a column that adds up to 10 from a 4x4
2018, only 9.2% of the credit offer is for more than two months’ credit,
including revolving payment (Japan Consumer Credit Association Statistics). decimal matrix. This question is incentivized in the way that re
The reluctance to use revolving payment and the popularity of monthly clear is spondents obtain more bonus points that are exchangeable for gift cer
possibly due to the preference of Japanese consumers and the default settings of tificates by answering quickly. The participants are asked to complete
credit cards. five tasks in 300 seconds. The remaining time is reduced by 60 seconds
5
for each incorrect answer. After the completion of the tasks, the Table 3
remaining seconds ×0.1 points are added to the participants’ bonus Summary Statistics
points. While calculation ability is considered to be a part of financial count mean sd min max
literacy in some existing studies (See Gerardi et al. (2013), for example),
Speculative Investment 5848 0.07 0.19 0 1
we use this variable to isolate pure calculation ability from numeracy in Share of Risky Assets 5848 0.12 0.23 0 1
financial situations captured by questions FL1 and FL2. Overborrowing 5848 0.10 0.18 0 1
Financial Naivety 5848 0.31 0.14 0 1
3.2.3. Preferences and beliefs Insufficient Retirement Planning 5848 0.78 0.21 0 1
Gambling Interest 5848 0.68 0.40 0 1
Preference and belief variables include confidence gap in financial Financial Literacy 5848 0.54 0.28 0 1
literacy, discount rate, present bias, risk aversion, loss aversion, S-shaped Financial Education University 5848 0.06 0.24 0 1
probability weighting, and ambiguity aversion. School Education 5845 2.12 1.07 0 5
Confidence gap in financial literacy is the respondent’s expectation of Financial Experience 5848 0.06 0.17 0 1
Calculation Time (Seconds) 5847 172 160 1 2477
his Financial literacy (obtained from question CFL in Table 2) minus his
Confidence Gap in Financial Literacy 5848 -0.10 0.22 -1 1
actual Financial literacy. Discount Rate 5848 0.13 0.11 0.00 0.30
Discount rate is a one-year discount rate of 10,000 hypothetical yen Present Bias 5848 0.00 0.06 -0.30 0.30
(approximately 90 dollars) obtained by the switching point of a price list Risk Aversion 5051 0.36 0.82 -1.02 1.35
question.5 Loss Aversion 5848 6.30 7.33 0.1 20
S-Shaped Probability Weighting 5848 0.39 0.24 -0.45 0.95
Present bias is Discount rate minus the discount rate for the period of Ambiguity Aversion 5848 0.81 0.39 0 1
one month to one year and one month after. Male 5848 0.50 0.50 0 1
Risk aversion is the absolute risk aversion obtained by the switching Married 5848 0.62 0.49 0 1
point of a price list of multiple pairs of lotteries. They are incentivized by Household Size 5848 2.77 1.28 1 7
Household Income (Ten Thousand 5766 566 378 150 2000
reward points that are exchangeable for gift certificates. The re
Yen)
spondents are asked to select one from two possible lotteries (safe lottery Home Ownership 5848 0.72 0.45 0 1
and risky lottery) in 10 cases (cases 1–10) in the same manner as in Holt Population Density (Thousand per 5843 3.87 4.69 0.00 22.38
and Laury (2002). In case X∈ {1, 2, ⋯, 10}, the participants earn 40 Square Kilometer)
points (10 × X%) or 32 points (100 − 10 × X%) in the safe lottery, and Observations 5848
77 points (10 × X%) or 2 points (100 − 10 × X%) in the risky lottery.

One of the cases is randomly selected, and the result in the chosen lottery gender variables, are excluded from the regression results tables.
is added to the bonus points.
Loss aversion is the marginal loss to gain ratio for participation in a 3.2.5. Descriptive Statistics
hypothetical coin flip question LA in Table 2, gaining a certain amount The summary statistics for the variables used in the regressions are
of money for heads and losing 10,000 yen (approximately 90 dollars) for presented in Table 3. The Table shows that the average Japanese
tails. household is underconfident by a confidence gap of -0.10.7 Table 4
S-shaped probability weighting is the tendency to overweight small shows the correlation matrix for the knowledge and skills variables, and
probabilities and to underweight larger ones calculated from the re preference and belief variables. Financial literacy is positively correlated
sponses to the three hypothetical lottery questions from Tanaka et al. with other knowledge and skills variables, risk aversion, and ambiguity
(2010). The switching points in the price lists of lotteries is used to aversion, and is negatively correlated with confidence gap in financial
calculate S-shaped probability weighting. literacy and discount rate.
Ambiguity aversion is the degree to which ambiguous choices are To provide an indication of how the sample differs from other
avoided. The question used to calculate the variable is AA in Table 2. studies, we present the percentage of correct answers to the inflation
Ambiguity aversion is the dummy variable for the known-ratio urn choice question (FL3) below. The number in our sample is 65% compared to
in the hypothetical Ellsberg urn. 59% in 2010 as per the Japan Household Panel Survey on Consumer
Preferences and Satisfaction of Osaka University8 and 56% in 2016 and
3.2.4. Other individual characteristics 55% in 2019 regarding the Financial Literacy Survey of the Central
Other individual variables include age groups (30s, 40s, 50s, 60s, 70 Council for Financial Services Information of Japan. In the U.S., the
and over), gender (male), marital status, household size, household income, number was 65% in 2009 according to the Health and Retirement
home ownership, and population density. Study.9
Household income is log income obtained from the income ranks of
less than 300, 300 to 499, 500 to 699, 700 to 999, 1000 to 1499, and
3.3. Age and gender
1500 or more in tens of thousands of yen. Income is set to be equal to
150, 400, 600, 850, 1250, and 2000, respectively.
This subsection explores the relationship between age, gender, and
Population density refers to log population density of the respondent’s
the determinants of financial behaviors and attitudes using binned
residential municipality.
scatterplot (binscatter) to examine the implications for financial edu
Net household asset is not introduced in our regressions because the
cation. The results are presented in Figure 1. In all six scatter plots in
response rate to the question is very low. We assume net household asset
Figure 1, the horizontal axis is age. The vertical axes in the first and
is proxied by household income and age under the life-cycle hypothesis.6
second columns are financial literacy and financial experience in the first
The estimated coefficients for these variables, except for the age and
row, calculation ability and confidence gap in financial literacy in the
5 7
Respondents are asked to choose between a hypothetical option A, which It is possible that the expectations of the scores are underestimated by the
currently receives 10,000 yen, and a hypothetical option B, which receives fear of overestimation that may cause respondents embarrassment. We believe,
10,000 yen +α in a year’s time. α is presented in a ten-level price list ranging however, that this subject-expectancy effect is minimized when using a web-
from 0 to 3,000 yen. As α increases, respondents will change their choice from A based survey.
8
to B at some point. This is the switching point of the price list. The survey is also referred to as Preference Parameters Study or Survey of
6
Response rate to the income question is 100% because it is introduced as a Living Preferences and Satisfaction.
9
default question to all surveys conducted by the web survey company. See Hastings et al. (2013) for other countries.
6
second row, and discount rate and risk aversion in the third row. Results
using male samples are shown in blue ink and those using female sam
ples are shown in red ink. We will call each scatter plot as ”X panel”,
Ambiguity
Aversion
where X is the name of the variable on the vertical axis hereafter. Re
1.00
spondents are grouped into 12 equal-sized bins according to their age.
Quadratic fit lines and 95% confidence intervals are plotted along with
binned scatterplots.
The financial literacy panel shows that people become financially
more literate almost linearly with aging and that males are financially
Probability
Weighting
S-Shaped
more literate than females on average.10 This age effect contrasts with
0.03*
the negative age effect on calculation ability shown in the calculation
1.00
ability panel. Financial experience, which has similar effects as financial
literacy, also increases with age, and is higher for females than for males
Aversion
on average.
The confidence gap in financial literacy panel demonstrates that
1.00
0.02
0.02
Loss
people (who are mostly underconfident) become more unconfident in

financial literacy as they age and that females are more unconfident than
Aversion
males on average.
0.13***
0.07***
0.18*** The discount rate panel in Figure 1 shows that the point estimates of
1.00
Risk
this rate is highest in the 60s and early 70s, and it is higher for males
than for females on average until around 60 years old. The risk aversion
Present
panel in Figure 1 indicates that males are less risk-averse than females in
1.00
0.02
0.02
0.02
0.01
Bias
the 40s to early 70s on average.

0.29***
0.05***
Discout
3.4. Econometric framework

0.03*
0.03*
Rate
1.00
0.02
Our hypothesis is that financial literacy has both positive and

negative effects on financial behaviors and attitudes. Although it is
Confidence Gap in
commonly believed that higher levels of financial literacy induce

Financial Literacy
appropriate financial behaviors and attitudes, inappropriate ones can

also be induced as we saw in the several existing experimental studies in
-0.10***
-0.08***
Section 2. Bianchi (2018) points out that financial literacy leads to

-0.01
-0.02
1.00
0.02
0.01
holding riskier positions and rebalancing portfolios more frequently. As

for the confidence gap in financial literacy, we predict that it has
Calculation
negative effects on financial behaviors. Anderson et al. (2017) stress the

0.06***
0.04***
importance of perceived, not actual, literacy.

-0.01
1.00
0.01
0.01
0.01
0.01
Our hypothesis about the effect of discount rate and present bias is
that they have negative effects on speculative investment and share of
risky assets, while also having positive effects on overborrowing and
insufficient retirement planning. As for the effect of risk aversion, loss
Experience
Financial
aversion, and ambiguity aversion, we predict that they have negative

-0.03*
-0.02
1.00
0.02
0.01
0.00
0.01
0.01
0.00
effects on all financial behaviors and attitudes, as these factors work to

contain the risk-taking that is needed for more financial activities (or
inactivity for retirement planning).
Education
To test our hypotheses, we estimate the effects of financial literacy,

-0.08***
-0.05***
0.12***
0.04**
0.04**
School
preferences, and beliefs on financial behaviors and attitudes using the

-0.02
1.00
0.00
0.01
0.02
following regression model:

Yi = Ki β0 + Pi β1 + Xi γ + α + ϵi , (1)
where Yi is a vector of the financial behaviors and attitudes of respon

Education at
University
dent i; Ki is a vector of the knowledge and skills, including financial lit

Financial
0.19***
0.11***
0.07***
eracy; Pi is a vector of preference and belief variables; Xi is a vector of

Correlation Matrix for Independent Variables
0.03*
-0.01
-0.02
1.00
0.00
0.01
0.02
0.01
other individual characteristics; α denotes the constant term; and ϵi is an

error term. All the dependent and independent variables are standard
ized. We will report the results of estimating Equation (1) by OLS as our
Financial
-0.44***
-0.06***
baseline results in Section 4.1. In Section 4.2, we estimate using the

Literacy
0.12***
0.18***
0.12***
0.15***
0.23***
0.17***
breakdown of financial literacy test results. In Section 4.3, we conduct

1.00
0.00
0.00
0.02
robustness checks on the choice of independent variables using parsi

monious regressions by OLS, then move on to robustness checks on the
S-Shaped Probability
Financial Experience
Financial Literacy
Ambiguity Aversion
Financial Education
choice of estimation methods using ordered probit estimation and IV

Calculation Ability
Confidence Gap in
Financial Literacy
School Education
estimation.
at University
Risk Aversion
Loss Aversion
Discout Rate
Weighting
Present Bias
Table 4
10
The linear age effect is consistent with Sanchez and Dunning (2018), who
analyze a nationally representative sample of American adults.
7
Fig. 1. Age, Gender, and the Determinants of Financial Behaviors and Attitudes: OLS
To clarify the relationship between the independent variables be discount rates is consistent with Bianchi (2018), its estimates of the ef
forehand, financial literacy, our key independent variable, is regressed on fect of risk aversion and ambiguity aversion are insignificant.
other independent variables in Equation (1). The result in Table 5 in The negative coefficient of confidence gap in financial literacy implies
dicates that financial literacy is closely correlated with other knowledge the existence of a sort-of Dunning-Kruger effect, that is, incompetent
and skills variables, preference and belief variables, and other individual people do not have the ability to realize their own incompetence.11 This
characteristics. The signs of the estimated coefficients are the same as result is consistent with the finding of Anderson et al. (2017) that the
those of the correlation coefficients presented in Table 4. Greater Dunning-Kruger effect exists in the context of financial literacy.
financial literacy is attained by those who are educated, more financially The estimated effect of discount rate is consistent with Meier and
experienced, fast in calculation, unconfident in financial literacy, low on Sprenger (2013), who show that individuals choosing to acquire finan
discount rate, risk-averse, with S-shaped probability weighting, ambi cial literacy discount the future less.
guity averse, and older. While the sign of the estimated effect of low
11
Our simple estimate cannot escape the criticism that other factors, such as
”regression to the mean”, may be at play. For the possible different accounts of
the Dunning-Kruger effect, see Schlösser et al. (2013)
8
Table 5 literacy. The regression result, however, shows that respondents with
Financial Literacy Regression: OLS high levels of financial literacy tend to engage in more naive behavior
Financial Literacy related to financial markets.
Combining the results in columns (1)-(4), we can summarize that
Financial Education at University 0.046***
(0.010) people with higher levels of financial literacy are “daring and reckless”;
School Education 0.119*** that is, they take risks, overborrow, and behave in a more naive manner
(0.012) financially.
Financial Experience 0.082*** In contrast to the above regression results, the ones in columns (5)
(0.011)
Calculation Ability 0.130***
and (6) show that high levels of financial literacy lead to less insufficient
(0.012) retirement planning and gambling interest. These results are consistent with
Confidence Gap in Financial Literacy -0.410*** the conventional view that higher levels of financial literacy prevent
(0.010) inappropriate financial behaviors and attitudes. They are also consistent
Discount Rate -0.057***
with Duflo and Saez (2003), Sekita (2011), and Watanapongvanich
(0.012)
Present Bias 0.023* et al. (2020). One may notice that financial literacy has the opposite sign
(0.012) in the speculative investment and the gambling interest regressions. The
Risk Aversion 0.165*** respondents do not seem to comprehend gambling as a form of specu
(0.011) lative investment, while financially literate people are discreet about
Loss Aversion -0.010
(0.012)
this activity.12
S-Shaped Probability Weighting 0.023** The greater confidence gap in financial literacy has positive effects on
(0.011) speculative investment, share of risky assets, financial naivety, and has
Ambiguity Aversion 0.085*** negative effect on insufficient retirement planning. These results are
(0.012)
similar to the effect of financial literacy and suggest that individual be
30s 0.016
(0.016) haviors can be determined not only by actual literacy but also by
40s 0.093*** “perceived” literacy.
(0.017) Anderson et al. (2017) find that the effect of financial literacy on
50s 0.100*** retirement planning is mostly driven by perceived literacy, not by actual
(0.017)
literacy. In our analyses, however, the standardized negative coefficient
60s 0.198***
(0.018) for actual financial literacy in the retirement planning regression nearly
70 and over 0.218*** triples that of confidence gap in financial literacy, as is shown in column
(0.018) (5) of Table 6, which is certainly an encouraging result for financial
Male 0.148***
economics educators. The difference between the results of Anderson
(0.012)
R-Squared 0.391 et al. (2017) and this study warrants further investigation. Note, how
N 4968 ever, that the dominance of financial literacy is observed in other re
gressions in Table 6 with positive coefficients for both financial literacy
Robust standard errors in parentheses
and confidence gap in financial literacy. These results suggest that actual
* p < 0.1, ** p < 0.05, *** p < 0.01
The regression also includes marital status, household size, household income, financial literacy can induce reckless and prudent behaviors more than
home ownership, population density, and a constant term. the confidence gap (overconfidence and underconfidence) in financial
literacy.
4. Results Financial education at university does not seem to have a significant
impact on financial behaviors and attitudes, in contrast to its significant
4.1. Benchmark regressions effect on financial literacy, as presented in Table 5. School education
contributes to prevent overborrowing and insufficient retirement planning.
Table 6 shows the baseline estimates of Equation (1) using OLS. The It may induce, however, taking more risks; a result consistent with Black
coefficients for several independent variables (marital status, household et al. (2018). The effect on overborrowing opposes that of financial liter
size, household income, home ownership, and population density) are acy, suggesting that what is crucial for preventing overborrowing is not
excluded from the table for brevity purposes but are available upon financial literacy but school education. The effects of financial experience
request. Columns (1) and (2) show the results for speculative investment are similar to those of financial literacy, except for the effects on financial
and share of risky assets. Financial literacy has positive and significant naivety and gambling interest.
effects in both regressions. The result demonstrates that people with Calculation ability has no significant impact, except for the weak
high levels of financial literacy are highly active in financial markets, impact on financial naivety at the 10% significance level, which may
but the result showing the active involvement in speculative investment is seem to disagree with Gerardi et al. (2013), who show that people with
a warning sign. The result for share of risky assets is consistent with van high numerical abilities tend to avoid defaulting on their mortgage.
Rooij et al. (2011). Their measure of calculation ability is, however, very different from
The regression results in columns (3) and (4) show the results for ours. While our calculation ability is the addition speed of decimal
overborrowing and financial naivety. Overborrowing means that the numbers, Gerardi et al. (2013) use the score of calculation problems in
respondent has borrowed money at a very high-interest rate or has had everyday life situations, including interest rate calculation. Our results
difficulty with debt repayment. The positive coefficient for financial lit
eracy implies that financially literate respondents tend to have experi
enced being too deep in debt. The result is consistent with Bruhn et al. 12
The correlation between speculative investment and gambling interest is,
(2016) and inconsistent with the short term result of Skimmyhorn however, not significantly different from zero (the group with 80% or higher
(2016). Note that the coefficient for financial literacy has the opposite correct answers to the financial literacy quiz had a positive and significant
sign of that for school education. The contrasting result suggests that correlation, but the correlation coefficient was 0.06). This result is not incon
financial literacy can have specific financial behavioral influences, sistent with the results in Table 6, as financial literacy is not the only factor that
different from those of general education. Financial naivety reflects determines financial behaviors. For example, risk aversion suppresses both
inadvertent attitudes, which seemingly contradict high financial speculative investment and gambling interest. Thus, those who engage in specu
lative investment behavior are not necessarily uninterested in gambling.
9
Table 6
Financial Behavior: Baseline
(1) (2) (3) (4) (5) (6)
Speculative Share of Risky Over- Financial Insufficient Retirement Gambling

Investment Assets borrowing Naivety Planning Interest
Financial Literacy 0.234*** 0.136*** 0.112*** 0.148*** -0.318*** -0.165***

(0.018) (0.018) (0.017) (0.016) (0.016) (0.017)
Financial Education at University -0.018 0.017 0.003 0.005 -0.021 -0.005
(0.016) (0.016) (0.014) (0.016) (0.015) (0.014)
School Education -0.009 0.085*** -0.095*** -0.008 -0.030** -0.026*
(0.016) (0.016) (0.015) (0.016) (0.014) (0.015)
Financial Experience 0.062*** 0.056*** 0.038** 0.013 -0.073*** 0.027*
(0.018) (0.016) (0.015) (0.016) (0.014) (0.014)
Calculation Ability -0.010 -0.015 -0.003 -0.024* 0.006 0.001
(0.014) (0.014) (0.014) (0.014) (0.013) (0.014)
Confidence Gap in Financial 0.102*** 0.097*** 0.016 0.039** -0.116*** -0.021
Literacy
(0.016) (0.016) (0.015) (0.016) (0.015) (0.016)
Discount Rate 0.026* -0.029* 0.125*** -0.050*** 0.103*** 0.044***
(0.015) (0.015) (0.016) (0.014) (0.013) (0.015)
Present Bias -0.029** 0.003 -0.024 0.006 -0.025* -0.003
(0.014) (0.019) (0.017) (0.015) (0.013) (0.013)
Risk Aversion -0.032** -0.069*** -0.028* 0.017 -0.030** -0.116***
(0.014) (0.015) (0.015) (0.014) (0.013) (0.014)
Loss Aversion -0.042*** -0.029** -0.038*** -0.037*** 0.016 -0.050***
(0.013) (0.014) (0.013) (0.013) (0.013) (0.014)
S-Shaped Probability Weighting 0.022 -0.030** 0.018 0.018 -0.000 -0.010
(0.014) (0.014) (0.014) (0.015) (0.013) (0.014)
Ambiguity Aversion 0.009 -0.041*** 0.013 0.027* -0.021* -0.062***
(0.014) (0.015) (0.013) (0.014) (0.012) (0.013)
30s 0.027 -0.049** 0.017 0.003 -0.038*** -0.003
(0.022) (0.019) (0.018) (0.018) (0.014) (0.017)
40s -0.049** -0.035 0.097*** 0.010 -0.055*** -0.001
(0.022) (0.022) (0.021) (0.019) (0.017) (0.019)
50s -0.073*** -0.051** 0.095*** 0.017 -0.121*** -0.071***
(0.022) (0.021) (0.021) (0.019) (0.017) (0.020)
60s -0.099*** -0.020 0.034 0.053** -0.250*** -0.112***
(0.024) (0.023) (0.022) (0.023) (0.020) (0.022)
70 and over -0.113*** 0.024 -0.006 0.090*** -0.199*** -0.137***
(0.025) (0.024) (0.022) (0.024) (0.020) (0.022)
Male 0.121*** 0.082*** 0.068*** -0.003 0.031** 0.111***
(0.015) (0.014) (0.014) (0.015) (0.013) (0.014)
R-Squared 0.082 0.082 0.074 0.035 0.230 0.100
N 4968 4968 4968 4968 4968 4968

* p < 0.1, ** p < 0.05, *** p < 0.01
All regressions also include marital status, household size, household income, home ownership, population density, and a constant term.
suggest that pure calculation ability unrelated to financial situations has in some financial aspects. Risk aversion and loss aversion have negative
no impact on financial behaviors and attitudes. effects as predicted on almost all financial aspects, with the exceptions of
The age variables captures age-related factors not included in other financial naivety for risk aversion and insufficient retirement planning for
independent variables in the regressions. Although interpreting the re loss aversion. S-shaped probability weighting is negatively related with
sults for the variable requires some speculations, we provide tentative share of risky assets, however, it is unrelated to gambling interest. Ambi
explanations for future research. Column (5) shows that retirement guity aversion has predicted negative effects on both share of risky assets
planning improves as people age. If we focus on the supply side of and gambling interest. The effect on share of risky assets is consistent with
financial opportunities, this result is understandable because people Dimmock et al. (2016), who show that ambiguity aversion is negatively
tend to have more opportunities to be exposed to information about correlated with stock market participation and the fraction of financial
retirement planning as they age. This age effect is consistent with Sekita assets in stocks.
(2020). Column (3) shows that overborrowing tends to be high in
middle-aged individuals. The abundant borrowing opportunities such as 4.2. Breakdown of financial literacy test
mortgages can be the cause of this middle-age effect. Financial naivety
accelerates for people in their sixties and seventies. During these ages To further investigate which aspects of financial literacy have led to
cognitive skills deteriorate. The age effect may arise from the cognitive the findings above, we estimate using the breakdown of financial liter
decline in the elderly not captured by our calculation variable. acy test results, that is, financial literacy is replaced with ten dummy
The coefficients of discount rate have an expected positive sign for variables, each of which corresponds to one of the ten financial literacy
overborrowing and insufficient retirement planning. The negative sign of questions. The OLS estimates are presented in Table 7. There are only a
financial naivety regression indicates that high discount rates suppress few cases in which specific financial literacy has the opposite effect to
naive financial behavior. This may reflect the fact that financially naive general financial literacy. It is encouraging that knowledge of risk-return
behaviors involve some forms of investment. The sign in gambling in tradeoff (FL10) curbs holding of risky assets. In columns (1)–(4), how
terest regression is positive, which may indicate that people with high ever, most aspects of financial knowledge lead to daring and reckless
discount rates are vulnerable to the get-rich-quick opportunities. financial behaviors. When it comes to financial naivety, financial literacy
Other behavioral factors tend to make people less daring and reckless is irrelevant at best, and some of it can have a negative impact. Financial
10
Table 7
Financial Behavior: Breakdown of Financial Literacy Test
(1) (2) (3) (4) (5) (6)

FL1 Interest 0.051*** 0.000 0.007 0.005 0.003 -0.062***

(0.016) (0.016) (0.017) (0.017) (0.015) (0.017)
FL2 Compound 0.048*** 0.051*** -0.004 0.012 -0.078*** -0.010
(0.017) (0.015) (0.017) (0.018) (0.015) (0.016)
FL3 Inflation 0.042** 0.011 0.028 0.020 -0.083*** -0.070***
(0.016) (0.017) (0.017) (0.018) (0.015) (0.017)
FL4 Bond Price 0.094*** 0.096*** -0.042*** -0.018 -0.047*** 0.003
(0.016) (0.015) (0.015) (0.016) (0.014) (0.015)
FL5 Mortgage 0.005 -0.019 0.062*** 0.034** -0.024 -0.029*
(0.017) (0.016) (0.017) (0.017) (0.016) (0.017)
FL6 Divers. 0.050*** 0.067*** -0.013 0.053*** -0.092*** -0.005
(0.016) (0.016) (0.018) (0.017) (0.015) (0.017)
FL7 Foreign 0.039** 0.049*** 0.020 0.012 -0.068*** -0.022
(0.016) (0.016) (0.016) (0.016) (0.015) (0.016)
FL8 Liquidity 0.050*** 0.058*** 0.040** 0.048*** -0.069*** -0.004
(0.017) (0.015) (0.016) (0.018) (0.015) (0.016)
FL9 Insurance 0.020 0.012 0.066*** 0.024 -0.011 0.015
(0.015) (0.014) (0.015) (0.015) (0.013) (0.014)
FL10 Risk-Return 0.012 -0.066*** 0.031* 0.046*** -0.050*** -0.061***
(0.018) (0.017) (0.018) (0.016) (0.015) (0.017)
Financial Education at University -0.019 0.016 0.004 0.006 -0.021 -0.005
(0.016) (0.016) (0.014) (0.016) (0.015) (0.014)
School Education -0.012 0.080*** -0.092*** -0.006 -0.028** -0.026*
(0.016) (0.016) (0.015) (0.016) (0.014) (0.015)
Financial Experience 0.060*** 0.051*** 0.041*** 0.015 -0.074*** 0.023
(0.019) (0.016) (0.015) (0.016) (0.014) (0.014)
Calculation Ability -0.011 -0.014 0.000 -0.020 0.004 0.003
(0.014) (0.014) (0.014) (0.014) (0.013) (0.014)
Literacy
(0.016) (0.017) (0.016) (0.016) (0.015) (0.016)
Discount Rate 0.027* -0.027* 0.124*** -0.051*** 0.103*** 0.044***
(0.015) (0.015) (0.016) (0.014) (0.013) (0.015)
Present Bias -0.029** 0.002 -0.025 0.005 -0.025* -0.006
(0.014) (0.019) (0.017) (0.015) (0.013) (0.013)
Risk Aversion -0.028* -0.057*** -0.031** 0.014 -0.031** -0.109***
(0.014) (0.015) (0.015) (0.014) (0.013) (0.014)
Loss Aversion -0.040*** -0.024* -0.038*** -0.036*** 0.015 -0.044***
(0.013) (0.014) (0.013) (0.013) (0.013) (0.014)
S-Shaped Probability Weighting 0.023 -0.028** 0.018 0.017 0.000 -0.007
(0.014) (0.014) (0.014) (0.015) (0.013) (0.014)
Ambiguity Aversion 0.012 -0.033** 0.013 0.025* -0.021* -0.056***
(0.014) (0.015) (0.013) (0.014) (0.012) (0.013)
R-Squared 0.086 0.097 0.082 0.039 0.236 0.106
N 4968 4968 4968 4968 4968 4968

* p < 0.1, ** p < 0.05, *** p < 0.01
All regressions also include age variables, male, marital status, household size, household income, home ownership,
population density, and a constant term.
naivety may be induced by knowledge about mortgages (FL5), diversi by robustness checks on the choice of independent variables using
fied investment (FL6), liquidity (FL8), and risk-return tradeoff (FL10), parsimonious regressions by OLS, then move on to robustness checks on
which can be summarized as financial knowledge required for personal the choice of estimation methods using ordered probit estimation and IV
investors. estimation.
Focusing on the questions (FL2, FL3, FL6) related to the Big Three, all First, we conduct parsimonious regressions to examine whether the
of them seem to induce more speculative investment, with two (FL2 and unintuitive sings of the parameter estimates of the coefficients on
FL6) leading to higher share of risky assets. In contrast, they are irrelevant financial literacy reported in Table 6 were statistical artifacts arising from
to overborrowing, and only one of them (FL6) leads to financial naivety. As the correlation with the other independent variables presented in
for overborrowing and financial naivety, daring and reckless behaviors are Table 4. In doing so we run two types of parsimonious regressions which
related to questions other than the Big Three, such as mortgage (FL5), exclude some of the independent variables used in Table 6. The results of
liquidity (FL8), insurance (FL9), and risk-return tradeoff (FL10). These estimations are reported in Tables 8 and 9. In Table 8, all knowledge,
results demonstrate that other types of financial knowledge which are skill, preference, and belief variables other than financial literacy are
not captured by the Big Three questions seem to be at play for over excluded from Equation (1). In Table 9, only knowledge and skill vari
borrowing and financial naivety. ables other than financial literacy are excluded. The estimated co
efficients for financial literacy have the same signs and significant at the
4.3. Robustness checks 1% level in Tables 6, 8, and 9. Despite the correlations between the
explanatory variables presented in Table 4, the estimated coefficients for
This subsection conducts two types of robustness checks. We begin financial literacy are robust to the exclusion of some of the explanatory
11
Table 8
Financial Behavior: Parsimonious (knowledge, skill, preference, and belief variables excluded)
(1) (2) (3) (4) (5) (6)
Speculative Investment Share of Risky Assets Over-borrowing Financial Naivety Insufficient Retirement Planning Gambling Interest

(0.013) (0.014) (0.013) (0.011) (0.012) (0.013)
30s 0.028 -0.060*** 0.030* 0.005 -0.037*** 0.000
(0.021) (0.019) (0.017) (0.016) (0.014) (0.017)
40s -0.044** -0.068*** 0.114*** 0.007 -0.057*** 0.004
(0.021) (0.021) (0.020) (0.018) (0.016) (0.018)
50s -0.069*** -0.076*** 0.119*** 0.021 -0.114*** -0.055***
(0.021) (0.020) (0.020) (0.018) (0.017) (0.018)
60s -0.086*** -0.048** 0.076*** 0.040* -0.235*** -0.096***
(0.022) (0.022) (0.021) (0.021) (0.018) (0.020)
70 and over -0.090*** -0.005 0.038* 0.078*** -0.188*** -0.104***
(0.022) (0.022) (0.021) (0.021) (0.018) (0.021)
Male 0.123*** 0.120*** 0.060*** -0.006 0.018 0.118***
(0.013) (0.013) (0.013) (0.013) (0.012) (0.013)
R-Squared 0.066 0.054 0.045 0.030 0.195 0.075
N 5761 5761 5761 5761 5761 5761

* p < 0.1, ** p < 0.05, *** p < 0.01
All regressions also include marital status, household size, household income, home ownership,
population density, and a constant term.
Table 9
Financial Behavior: Parsimonious (knowledge and skill variables excluded)
(1) (2) (3) (4) (5) (6)


(0.018) (0.018) (0.017) (0.015) (0.016) (0.017)
Literacy
(0.016) (0.016) (0.015) (0.016) (0.015) (0.016)
Discount Rate 0.026* -0.036** 0.131*** -0.050*** 0.107*** 0.045***
(0.015) (0.015) (0.016) (0.014) (0.014) (0.015)
Present Bias -0.028** 0.004 -0.023 0.006 -0.026* -0.003
(0.014) (0.019) (0.017) (0.015) (0.013) (0.013)
Risk Aversion -0.033** -0.070*** -0.029** 0.016 -0.030** -0.117***
(0.014) (0.015) (0.015) (0.014) (0.013) (0.014)
Loss Aversion -0.043*** -0.027* -0.042*** -0.037*** 0.016 -0.052***
(0.013) (0.014) (0.013) (0.013) (0.013) (0.014)
S-Shaped Probability Weighting 0.022 -0.034** 0.022 0.018 0.001 -0.008
(0.014) (0.014) (0.014) (0.015) (0.013) (0.014)
Ambiguity Aversion 0.007 -0.041*** 0.010 0.027* -0.020* -0.063***
(0.014) (0.015) (0.014) (0.014) (0.012) (0.013)
30s 0.027 -0.050** 0.017 0.003 -0.037** -0.003
(0.022) (0.020) (0.018) (0.018) (0.015) (0.017)
40s -0.045** -0.039* 0.105*** 0.012 -0.054*** 0.002
(0.022) (0.022) (0.021) (0.019) (0.017) (0.019)
50s -0.069*** -0.060*** 0.106*** 0.020 -0.118*** -0.067***
(0.022) (0.021) (0.021) (0.019) (0.017) (0.020)
60s -0.093*** -0.026 0.047** 0.057** -0.251*** -0.108***
(0.024) (0.023) (0.022) (0.023) (0.020) (0.022)
70 and over -0.107*** 0.010 0.014 0.096*** -0.194*** -0.131***
(0.024) (0.024) (0.022) (0.023) (0.019) (0.022)
Male 0.111*** 0.093*** 0.049*** -0.006 0.030** 0.104***
(0.014) (0.014) (0.014) (0.014) (0.013) (0.014)
R-Squared 0.078 0.072 0.065 0.034 0.223 0.099
N 4970 4970 4970 4970 4970 4970

* p < 0.1, ** p < 0.05, *** p < 0.01
All regressions also include marital status, household size, household income, home ownership, population density, and a constant term.
variables. Ordered probit estimation use the weak assumption that the six
Second, we conduct robustness checks on the choice of estimation financial behavior variables except for share of risky assets are ordinal.
methods. We report three results of the estimation of Equation 1 using Note that each of these variables is the share of questions with specific
different estimation methods: by ordered probit estimation, by IV esti answers within a group of questions. When the number of questions in a
mation, and by ordered probit IV estimation in Tables 10, A1, and A2, group is two, there are three possible values for the dependent variable:
respectively. 0, 1/2 and 1. While OLS estimate assumes that the distance of a
12
Table 10
Financial Behavior: Ordered Probit
(1) (2) (3) (4) (5)
Speculative Investment Over-borrowing Financial Naivety Insufficient Retirement Planning Gambling Interest
Financial Literacy 0.426*** 0.144*** 0.196*** -0.426*** -0.225***

(0.035) (0.023) (0.023) (0.021) (0.023)
Financial Education at University -0.026 0.020 0.005 -0.027* -0.009
(0.022) (0.017) (0.021) (0.016) (0.017)
School Education -0.009 -0.109*** -0.011 -0.037** -0.030
(0.026) (0.019) (0.021) (0.017) (0.019)
Financial Experience 0.086*** 0.043** 0.015 -0.084*** 0.032*
(0.022) (0.018) (0.020) (0.014) (0.018)
Calculation Ability -0.023 -0.009 -0.033* 0.007 0.002
(0.025) (0.018) (0.019) (0.015) (0.018)
Confidence Gap in Financial Literacy 0.164*** 0.023 0.048** -0.138*** -0.024
(0.027) (0.019) (0.022) (0.018) (0.020)
Discount Rate 0.042 0.138*** -0.051*** 0.123*** 0.052***
(0.026) (0.019) (0.020) (0.017) (0.019)
Present Bias -0.062** -0.042** 0.001 -0.031* -0.003
(0.028) (0.020) (0.020) (0.017) (0.018)
Risk Aversion -0.061** -0.033* 0.022 -0.036** -0.150***
(0.025) (0.018) (0.019) (0.016) (0.018)
Loss Aversion -0.070*** -0.047*** -0.051*** 0.029* -0.062***
(0.025) (0.018) (0.018) (0.016) (0.017)
S-Shaped Probability Weighting 0.046* 0.033* 0.023 0.005 -0.008
(0.024) (0.017) (0.020) (0.015) (0.017)
Ambiguity Aversion 0.014 0.018 0.045** -0.026* -0.089***
(0.026) (0.018) (0.020) (0.015) (0.018)
30s 0.043 0.013 0.003 -0.064*** -0.001
(0.033) (0.024) (0.025) (0.021) (0.024)
40s -0.069* 0.114*** 0.020 -0.089*** -0.004
(0.035) (0.026) (0.027) (0.023) (0.026)
50s -0.110*** 0.117*** 0.031 -0.174*** -0.093***
(0.037) (0.026) (0.027) (0.023) (0.025)
60s -0.152*** 0.050* 0.070** -0.312*** -0.141***
(0.040) (0.028) (0.031) (0.024) (0.027)
70 and over -0.178*** -0.023 0.113*** -0.264*** -0.172***
(0.042) (0.030) (0.031) (0.025) (0.028)
Male 0.207*** 0.067*** -0.021 0.069*** 0.142***
(0.027) (0.018) (0.019) (0.016) (0.018)
R-Squared 0.103 0.032 0.023 0.076 0.055
N 4968 4968 4968 4968 4968

* p < 0.1, ** p < 0.05, *** p < 0.01
All regressions also include marital status, household size, household income, home ownership, population density,
and a constant term.
dependent variable between 0 and 1/2 is equal to that between 1/2 and reasons behind the difference. First, due to endogeneity, OLS un
1. Ordered probit estimation relaxes this assumption. Instrumental derestimates the effect of financial literacy. Second, the local average
variable estimates aim to take into account the possible endogeneity of treatment effect (LATE) obtained by IV is larger than the average
financial literacy. The signs and significance of the coefficients of the treatment effect (ATE) obtained by OLS. Third, the exclusion restriction
ordered probit estimation results in Table 10 are not significantly is not met. Under the assumption that our instrumental variables are
different from those of the OLS results in Table 6. valid, the IV regression results confirm the significant effect of financial
The effect of financial literacy is the focus of our regression model. literacy.
This variable, however, could be endogenous, especially when the There are other differences between the two tables. Unlike the OLS
dependent variables are speculative investment, share of risky assets, and results, the IV results do not find that school education contributes to
overborrowing, because these financial behaviors may increase the re preventing insufficient retirement planning. With respect to share of risky
spondents’ exposure to financial knowledge and improve their financial assets, the effect of school education in the IV results is opposite to that of
literacy. To deal with this potential endogeneity problem, we employ OLS, with the result that it discourages this behavior. For calculation
two instrumental variables: parents’ financial work experience and ability the IV results show that this ability suppresses daring and reckless
parents’ stock trade experience. Those parents’ variables are assumed to financial behaviors, except for overborrowing, which is in contrast to the
be exogenous to our financial behavior equations. The F-statistic for insignificant OLS results and consistent with Gerardi et al. (2013).
testing weak instrument is 55, suggesting that the instruments are suf Minor differences between the OLS and the IV results include the
ficiently strong.13 following: for risk aversion, the coefficient for the insufficient retirement
The comparison between Tables 6 and A1 shows that the coefficients planning regression is positive for the IV results, which is contrary to the
on financial literacy is larger in IV regressions. There are several possible OLS results that risk aversion is negatively associated with daring and
reckless financial behaviors. For ambiguity aversion, its impact on insuf
ficient retirement planning was positive for IV, opposite to the OLS results.
13
The endogeneity test reject the null hypothesis that financial literacy is In the share of risky assets regression, the effects of discount rate and the
exogenous at the 5% level in all but overborrowing regression. The Hansen test gender variable are opposite to those of the OLS results. While financial
for overidentifying restrictions does not reject the null hypothesis that the in education at university does not have a significant effect on financial
struments are uncorrelated with the error term at the 5% level in all regressions.
13
Table A1
Financial Behavior: IV
(1) (2) (3) (4) (5) (6)

Financial Literacy 0.870*** 1.808*** 0.315** 0.530*** -0.943*** -0.535***

(0.164) (0.220) (0.128) (0.153) (0.148) (0.137)
Financial Education at University -0.047** -0.061** -0.006 -0.012 0.008 0.012
(0.019) (0.026) (0.015) (0.018) (0.018) (0.016)
School Education -0.085*** -0.114*** -0.119*** -0.054** 0.044* 0.018
(0.026) (0.037) (0.021) (0.024) (0.023) (0.022)
Financial Experience 0.010 -0.081*** 0.022 -0.019 -0.021 0.057***
(0.024) (0.030) (0.019) (0.022) (0.019) (0.018)
Calculation Ability -0.092*** -0.232*** -0.029 -0.074*** 0.087*** 0.049**
(0.026) (0.038) (0.023) (0.025) (0.024) (0.023)
Confidence Gap in Financial 0.362*** 0.781*** 0.100* 0.195*** -0.372*** -0.172***
Literacy
(0.069) (0.092) (0.054) (0.065) (0.062) (0.057)
Discount Rate 0.063*** 0.067** 0.137*** -0.028* 0.068*** 0.022
(0.019) (0.028) (0.017) (0.017) (0.017) (0.017)
Present Bias -0.044*** -0.035 -0.028* -0.003 -0.011 0.005
(0.017) (0.030) (0.017) (0.016) (0.015) (0.014)
Risk Aversion -0.137*** -0.344*** -0.062** -0.046 0.072*** -0.056**
(0.031) (0.043) (0.025) (0.029) (0.028) (0.027)
Loss Aversion -0.035** -0.011 -0.036*** -0.033** 0.010 -0.054***
(0.015) (0.025) (0.013) (0.014) (0.015) (0.015)
S-Shaped Probability Weighting 0.007 -0.068*** 0.013 0.009 0.014 -0.001
(0.016) (0.023) (0.014) (0.016) (0.015) (0.015)
Ambiguity Aversion -0.046** -0.184*** -0.004 -0.006 0.032* -0.030*
(0.021) (0.031) (0.017) (0.019) (0.019) (0.018)
30s 0.016 -0.076** 0.013 -0.004 -0.027 0.003
(0.024) (0.034) (0.018) (0.019) (0.018) (0.019)
40s -0.108*** -0.191*** 0.078*** -0.025 0.003 0.034
(0.029) (0.041) (0.025) (0.025) (0.024) (0.024)
50s -0.137*** -0.218*** 0.074*** -0.021 -0.059** -0.034
(0.029) (0.042) (0.025) (0.026) (0.025) (0.025)
60s -0.225*** -0.350*** -0.006 -0.023 -0.127*** -0.039
(0.041) (0.058) (0.035) (0.039) (0.037) (0.035)
70 and over -0.252*** -0.341*** -0.051 0.007 -0.062 -0.057
(0.045) (0.061) (0.037) (0.042) (0.039) (0.038)
Male 0.027 -0.165*** 0.038 -0.060** 0.123*** 0.165***
(0.029) (0.041) (0.023) (0.028) (0.026) (0.025)
N 4968 4968 4968 4968 4968 4968
Robust standard errors in parentheses * p < 0.1, ** p < 0.05, *** p < 0.01 All regressions also include marital status, household size, household income, home
ownership, population density, and a constant term.
behavior in the OLS results, the IV results show that it suppresses spec use their knowledge properly.
ulative investment behavior. The impact of confidence gap in financial lit To further explore the relationship between financial literacy and the
eracy on gambling interest is negative and significant in the IV estimation. financial behaviors and attitudes variable, we employ binscatter with
Thus, a comparison of the OLS and IV results does not necessarily semi-linear covariate-adjustment proposed by Cattaneo et al. (2019),
yield robust results for all variables and behaviors. The effects of school which correctly incorporates additional covariates under nonlinearity.14
education and calculation ability, in particular, need to be interpreted Figure 2 shows quadratic fit lines and 95% confidence intervals along
with caution. In contrast, there were no significant qualitative differ with binned scatterplots. The covariates we included in the estimates are
ences on the effects of financial literacy, loss aversion, and S-shaped the same as those of Table 6. In all six scatter plots in Figure 2, the
probability weighting. horizontal axis is financial literacy. The vertical axes in the first and
The estimates by the combination of the ordered probit and IV second columns are speculative investment and share of risky asset in the
(Table A2) also do not differ largely from the IV results in Table A1. The first row, overborrowing and financial naivety in the second row, and
result, combined with the similarity between Table 6 and Table 10 in insufficient retirement planning and gambling interest in the third row. We
dicates that our OLS and IV regression results are both robust to the will call each scatter plot as ”X panel”, where X is the name of the var
introduction of the ordered probit estimation method. iable on the vertical axis hereafter.
The panels in Figure 2 show almost no signs of nonlinearity leading
to the reversal of the literacy-behavior gradient, except for the financial
4.4. The literacy–behavior gradient naivety panel and the gambling interest panel at high financial literacy
levels. The observed reversal in the financial naivety panel is, however,
The regression results in Section 4.1 show that higher levels of very weak and far from offsetting the negative effect of financial liter
financial literacy are associated with some of the inappropriate financial acy. The negative effect of financial literacy, in contrast, accelerates in
behaviors and attitudes. A possible objection to this result states that the
effects of financial literacy on financial behaviors and attitudes are not
always negative. It is possible, for example, that the estimated effects
apply only to people with a low to moderate level of financial literacy; 14
The commonly-used residualized binscatter is inappropriate because Frisch-
that is, the effects could reverse in people with more than a certain level Waugh theorem does not apply under the nonlinear relationship between
of financial literacy, because financial experts are commonly believed to financial literacy and financial behaviors.
14
Table A2
Financial Behavior: IV Ordered Probit
(1) (2) (3) (4) (5)
Speculative Investment Over-borrowing Financial Naivety Insufficient Retirement Planning Gambling Interest
Financial Literacy 1.103*** 0.534*** 0.619*** -0.992*** -0.620***

(0.094) (0.140) (0.151) (0.092) (0.133)
Financial Education at University -0.057*** 0.001 -0.016 0.005 0.010
(0.020) (0.018) (0.022) (0.017) (0.018)
School Education -0.098*** -0.152*** -0.062** 0.042* 0.021
(0.027) (0.023) (0.027) (0.022) (0.026)
Financial Experience 0.007 0.008 -0.021 -0.023 0.064***
(0.026) (0.022) (0.024) (0.019) (0.020)
Calculation Ability -0.117*** -0.060** -0.087*** 0.086*** 0.055**
(0.026) (0.026) (0.027) (0.021) (0.025)
Confidence Gap in Financial Literacy 0.442*** 0.184*** 0.223*** -0.374*** -0.190***
(0.044) (0.061) (0.067) (0.042) (0.059)
Discount Rate 0.079*** 0.154*** -0.023 0.073*** 0.026
(0.023) (0.019) (0.022) (0.020) (0.022)
Present Bias -0.069*** -0.049*** -0.009 -0.013 0.007
(0.025) (0.019) (0.020) (0.016) (0.018)
Risk Aversion -0.175*** -0.096*** -0.052 0.070*** -0.075**
(0.027) (0.029) (0.033) (0.025) (0.034)
Loss Aversion -0.049** -0.041** -0.043** 0.019 -0.063***
(0.023) (0.018) (0.018) (0.016) (0.017)
S-Shaped Probability Weighting 0.021 0.022 0.012 0.018 0.001
(0.022) (0.018) (0.020) (0.015) (0.017)
Ambiguity Aversion -0.053** -0.017 0.006 0.030 -0.050**
(0.025) (0.021) (0.025) (0.018) (0.023)
30s 0.024 0.006 -0.004 -0.046** 0.006
(0.029) (0.024) (0.024) (0.021) (0.024)
40s -0.126*** 0.072** -0.022 -0.020 0.034
(0.032) (0.031) (0.031) (0.027) (0.029)
50s -0.164*** 0.072** -0.015 -0.092*** -0.048
(0.033) (0.032) (0.032) (0.029) (0.030)
60s -0.273*** -0.030 -0.019 -0.152*** -0.053
(0.037) (0.041) (0.045) (0.043) (0.043)
70 and over -0.306*** -0.108** 0.012 -0.098** -0.075
(0.039) (0.042) (0.050) (0.044) (0.046)
Male 0.057 0.005 -0.083*** 0.151*** 0.195***
(0.039) (0.029) (0.030) (0.020) (0.024)
N 4968 4968 4968 4968 4968

* p < 0.1, ** p < 0.05, *** p < 0.01
All regressions also include marital status, household size, household income, home ownership, population density,
and a constant term.
the speculative investment panel and the share of risky asset panel. The higher fees and using high-cost borrowing.” (p. 332). This statement
insufficient retirement planning panel indicates that the gradient is seems to contradict the regression results of this study, which show that
almost constant. This implies that financial literacy is effective in financial literacy is associated with high-cost borrowing or delinquency.
inducing people to prepare for their retirement, even for those people One possible reason for the difference is that factors related to financial
with the highest level of financial literacy. behaviors are employed comprehensively in this study due to the use of
our purpose-built survey data. Lusardi and Tufano (2015), for example,
5. Discussion and Conclusion does not consider education attainment, and proxied by income and
wealth. This omission of the education variable can cause a significant
We have estimated the effects of financial literacy and behavioral bias in the coefficient of financial literacy because financial literacy and
factors on the financial behavior and attitudes of Japanese households school education is positively correlated as presented in Table 4, and
and found that high financial literacy tends to be associated with spec financial literacy and school education have the opposite effects on
ulative investment, higher share of risky assets, overborrowing, and borrowing behavior as pointed out in Section 4.1.
financial naivety; that is, financial literacy makes people more daring Why does financial literacy lead to some inappropriate financial
and reckless financially. Moreover, there is no clear threshold above behaviors? Although providing an answer to the question is beyond the
which the effects of additional financial literacy are reversed. It is sur scope of this paper, five possible explanations are presented below:
prising that higher financial literacy at any level does not prevent overconfidence in financial ability, fluency effect, biased knowledge,
overborrowing and financial naivety, which are hardly considered sen cultural factors, and the level of financial literacy.
sible financial behaviors. First, higher levels of financial literacy may lead to increased confi
Holding more risky assets, which is not always inappropriate, can be dence in one’s financial ability that is not supported by an increase in
disastrous for consumers when these assets are combined with specu actual ability. Note that the confidence in financial ability, which is
lative investment, overborrowing, and financial naivety. These four different from the confidence in financial literacy, is very difficult to
types of financial behaviors and attitudes can all be induced, rather than measure and is not included in our analyses. Knecht (2013) argues, in
prevented, by higher levels of financial literacy. the context of aviation accidents, that although a linear relation is
Lusardi and Tufano (2015) conclude that “individuals with lower typically assumed between a pilot’s total flight hours and aviation ac
levels of debt literacy tend to transact in high-cost manners, incurring cident rate, it is actually a nonlinear relation; that is, fatalities occur
15
most frequently during the middle ranges of total flight hours. The knowledge would lead to financial recklessness. Considering the fact
relation is “perhaps because pilots are overconfident” (p.50). In the that research on the relationship between culture and decision-making
context of learning-by-doing, Sanchez and Dunning (2018) investigate seems to be still limited,16 further research is required in order to
six studies including Knecht (2013), and point out that beginners tend to clarify the effect of cultural factors.
be overconfident about their performance before reaching an interme Fifth, the expected level of background knowledge to answer our
diate level. As for the confidence of people who acquired additional financial literacy question can be the cause of the unintuitive result. As
expertise, posing industry-specific questions to IT professionals and UC the questions are designed to capture the financial literacy of the general
San Diego-specific questions to UCSD students, Mckenzie et al. (2008) population, their level is quite rudimentary for the knowledge of pro
show that experts and novices are about equally overconfident. fessional financiers. Evaluation of more sophisticated investor behaviors
These observations suggest that increased knowledge and skills can using an intermediate level financial quiz could lead to different con
affect confidence in various ways. Kahneman and Klein (2009) indicate clusions. Still, it is possible that people with the highest levels of
that professional intuition is sometimes overconfident and biased, and financial literacy are not good at making financial decisions, as illus
conclude that ”evaluating the likely quality of an intuitive judgment trated, for example, by the collapse of Long-Term Capital Management
requires an assessment of the predictability of the environment in which in 2000.
the judgment is made and of the individual’s opportunity to learn the As Charlie Munger, the co-chair of Berkshire Hathaway, once said,
regularities of that environment” (p.515). The advantage of this intui “There must be some wisdom in the folk saying: It’s the strong swimmers
tive expertise theory is that it can explain the overconfidence in specu who drown.”17 This mindset is consistent with our results, showing that
lative investment and the absence of overconfidence in retirement financial literacy leads people to engage in daring and reckless financial
planning at the same time: experts in finance become overconfident behaviors and attitudes. Our interpretation of the aphorism is that
under unpredictable financial situations, such as speculative investment, strong swimmers are overconfident in their swimming ability, feel
but, in contrast, are prudent under predictable financial situations, such familiar in dangerous waters (fluency effect), and know nothing about
as retirement planning. ocean currents (biased knowledge). This is especially noteworthy
Second, financial literacy may cause daring and reckless behaviors because financial literacy education is considered a key tool by policy
through the ”fluency effect.” Alter and Oppenheimer (2006) found that makers to protect ordinary people, including innocent young and
fluently named stocks outperform stocks with disfluent names. They affluent elderly people, from speculation or fraud. If financial literacy
point out that ”fluency gives rise to feelings of familiarity and a positive can lead to some inappropriate financial behaviors, what is important is
affective response” (p.9369). Financial fluency may also contribute to to teach the ”strong swimmer effect” along with general financial
people’s familiarity with financial products. In one of our financial knowledge.
naivety questions, we ask respondents if they ”feel attracted to casualty When it comes to biased knowledge, we are in dire need of more
insurance that offers compensation even for small damages.” It is specific, down-to-earth education. Food education, for example, teaches
possible that financially literate people are more fluent in reading and that unhealthy foods are plentiful on supermarket shelves. This is not the
understanding the description and felt familiar with financial terms like case for financial literacy education in Japan, as illustrated by The
”casualty insurance” and ”compensation.” Council for the Promotion of Financial and Economic Education (2016).
The advantage of the second explanation is that it can clarify the In fact, a number of financial products are inappropriate or unnecessary
puzzling result: financial literacy leads to both recklessness and pru for the average consumer. Buying individual stocks, for example, is a
dence depending on the type of financial behaviors. Under the fluency common investment vehicle, but it is extremely risky and requires
effect, people who are familiar with finance may prefer any activity professional-level financial knowledge and a lot of assets for diversifi
related to finance: speculative investment, risky assets, borrowing op cation. The true cost of mutual funds is often extremely hard to tell (The
portunities, financial advertisements, and retirement planning. They Nikkei (13 December, 2017)). Their cost transparency is extremely low
tend to be financially active irrespective of whether the activity is compared to the fat content of food items. There are many insurance
considered reckless or prudent. This explains the mixture of recklessness products that cover small losses, but it is not easy to identify these
and prudence of financially literate people. products that are being disadvantageous to consumers. There are major
Third, financial knowledge, measured by commonly used financial credit card companies in Japan that solicit consumers without clearly
literacy questions, is biased toward the perfect financial market. The Big indicating that the card is a revolving payment only card (The Nikkei (24
Three questions (about interest rate calculation, inflation, and diversi April, 2020)). Although The Council for the Promotion of Financial and
fied investment) do not include any knowledge about asymmetric in Economic Education (2016) warns against safe, high-return investments
formation, imperfect competition, and irrational behaviors, under and the use of revolving payments, the importance of vigilance against
which naive consumers can easily be taken advantage of by greedy fi inappropriate or unnecessary financial products does not seem to be
nanciers. In reality, it is crucial for consumers to know that they may be emphasized enough. To put these examples into the perspective of
corralled by companies with an information advantage, being forced to financial and economic literacy, we need to educate consumers that they
choose from a limited number of seemingly advantageous financial are in constant danger of being preyed upon by the presence of behav
products.15 ioral bias and information asymmetry and that such danger is particu
Fourth, cultural factors may have affected the estimation results. As larly high in the financial services industry. Consumers are swimming in
for overconfidence, Yates et al. (2010) find that Japanese respondents a sea of financial products, and being a strong swimmer can backfire and
are underconfident on average in terms of general knowledge questions, cause you to drown.
significantly unlike American and Chinese respondents, who are over Our results are also significant in considering financial fraud. fin
confident on average. This cultural difference suggests that there is (2013) points out that “older Americans are particularly vulnerable” (p.
considerable room for confidence building among Japanese consumers. 3) and that “males are more willing to take on investment risk than fe
One might hypothesize that this latitude would increase the impact of males” (p. 6). Notice that, as Figure 1 shows, elderly males are not as
increased financial knowledge on Japanese confidence in their own
financial behaviors, making it more likely that increased financial
16
The study by Yates and de Oliveira (2016) is an exceptions.
17
The same saying exists in Japan: ”Yoku oyogu mono ha oboru.”
15
Fujiki (2020a) points out that Japanese households’ sources of financial
knowledge are more skewed toward financial institutions and financial experts
than they would like.
16
Fig. 2. Effects of Financial Literacy on Financial Behavior: Binscatter
risk-loving as is suggested by the report. Rather, their high level of causal inference. Instrumental variable estimation is not a perfect so
financial knowledge can be a contributing factor to financial fraud lution. Existing studies that use a strict causal inference framework focus
vulnerability through daring and reckless behaviors and attitudes.18 The mostly on the effects of small changes in financial literacy in a specific
figure shows that males in their late sixties and seventies achieve the group of people. Future research should explore the causal effect of
highest financial quiz score. Note also that males in their 60s tend to larger financial literacy changes on financial behaviors using data from
have higher discount rates. Table 6 shows that high discount rates can the general population. Second, our measure of financial behaviors and
lead to overborrowing, insufficient retirement planning, and gambling attitudes is based on the answers to our survey questions. Analyses based
interest. on data from actual financial transactions are required. Third, the
While this study demonstrates that financial literacy can lead to method of teaching finance without incurring reckless behaviors has
inappropriate financial behaviors and attitudes, several issues warrant practical usefulness. Fourth, we need analysis based on surveys with
further research. First, our main results are not derived from a rigorous teenage respondents. Our results show that age is an important deter
minant of financial behaviors and attitudes. Therefore, it is entirely
possible that teenagers behave differently from people aged 20 or above
18
in our sample. Moreover, it is often overlooked that teenagers are facing
Using survey data from Japan, Yamori and Ueyama (2018) shows that
their own financial problems, such as student loans or online gambling.
people who have been scammed have higher average financial knowledge than
Financial education based on evidence from surveys on teenagers is,
those who have not.
17
therefore, required. Fouarge, D., Kriechel, B., Dohmen, T., 2014. Occupational sorting of school graduates:
The role of economic preferences. Journal of Economic Behavior and Organization
106, 335–351. https://doi.org/10.1016/j.jebo.2014.07.007.
Fujiki, H., 2020. Are the actual and intended sources of financial knowledge the same?
Declaration of Competing Interest evidence from japan. Japan and the World Economy 55, 101026.
Fujiki, H., 2020. Cash demand and financial literacy: A case study using japanese survey
data. Japan and the World Economy 100998.
None. Fujiki, H., 2020. Who adopts crypto assets in japan? evidence from the 2019 financial
literacy survey. Journal of the Japanese and International Economies 58, 101107.
Gerardi, K., Goette, L., Meier, S., 2013. Numerical ability predicts mortgage default.
Appendix A. Appendix Proceedings of the National Academy of Sciences of the United States of America 110
(28), 11267–11271. https://doi.org/10.1073/pnas.1220568110.
Goldzahl, L., 2017. Contributions of risk preference, time orientation and perceptions to
breast cancer screening regularity. Social Science and Medicine 185, 147–157.
https://doi.org/10.1016/j.socscimed.2017.04.037.
References Grinblatt, M., Keloharju, M., 2009. Sensation seeking, overconfidence, and trading
activity. Journal of Finance 64 (2), 549–578. https://doi.org/10.1111/j.1540-
6261.2009.01443.x.
, 2013Financial Fraud and Fraud Susceptibility in the United States, Technical Report,
Guiso, L., Paiella, M., 2008. Risk Aversion, Wealth, and Background Risk. Journal of the
2013.
European Economic Association 6 (6), 1109–1150.
Allgood, S., Walstad, W.B., 2016. The effects of perceived and actual financial literacy on
Hastings, J.S., Madrian, B.C., Skimmyhorn, B., 2013. Financial Literacy, Financial
financial behaviors. Economic Inquiry 54 (1), 675–697. https://doi.org/10.1111/
Education, and Economic Outcomes. Ssrn 347–375. https://doi.org/10.1146/
ecin.12255.
annurev-economics-082312-125807.
Alter, A.L., Oppenheimer, D.M., 2006. Predicting short-term stock fluctuations by using
Holt, C.A., Laury, S.K., 2002. Risk Aversion and Incentive Effects. American Economic
processing fluency, 103 (24), 9369–9372.
Review 92 (5), 1644–1655.
Anderson, A., Baker, F., Robinson, D.T., 2017. Precautionary savings, retirement
Horton, J.J., Rand, D.G., Zeckhauser, R.J., 2011. The online laboratory: Conducting
planning and misperceptions of financial literacy. Journal of Financial Economics
experiments in a real labor market. Experimental Economics 14 (3), 399–425.
126 (2), 383–398. https://doi.org/10.1016/j.jfineco.2017.07.008.
https://doi.org/10.1007/s10683-011-9273-9.
Anderson, L.R., Mellor, J.M., 2008. Predicting health behaviors with an experimental
Kadoya, Y., Khan, M., 2017. Explaining Financial Literacy in Japan: New Evidence Using
measure of risk preference. Journal of Health Economics 27 (5), 1260–1274. https://
Financial Knowledge, Behavior, and Attitude. SSRN Electronic Journal 1–27.
doi.org/10.1016/j.jhealeco.2008.05.011.
https://doi.org/10.2139/ssrn.3067799.
Arechar, A.A., Gächter, S., Molleman, L., 2018. Conducting interactive experiments
Kadoya, Y., Khan, M., Rabbani, N., 2017. Does Financial Literacy Affect Stock Market
online. Experimental Economics 21 (1), 99–131. https://doi.org/10.1007/s10683-
Participation? SSRN Electronic Journal (15), 1–2. https://doi.org/10.2139/
017-9527-2.
ssrn.3056562.
Barsky, R.B., Juster, F.T., Kimball, M.S., Shapiro, M.D., 1997. Preference Parameters and
Kadoya, Y., Khan, M.S.R., 2020. What determines financial literacy in Japan. Journal of
Behavioral Heterogeneity: an Experimental Approach in the Health and Retirement
Pension Economics and Finance 19 (3), 353–371. https://doi.org/10.1017/
Study. Quarterly Journal of Economics 112 (2), 537–579.
S1474747218000379.
Beshears, J., Choi, J.J., Laibson, D., Madrian, B.C., 2018. Behavioral Household Finance.
Kahneman, D., Klein, G., 2009. Conditions for Intuitive Expertise: A Failure to Disagree.
Handbook of Behavioral Economics, Volume 1. Elsevier B.V., pp. 177–276. https://
American Psychologist 64 (6), 515–526. https://doi.org/10.1037/a0016755.
doi.org/10.1016/bs.hesbe.2018.07.004
Kinari, Y., 2016. Properties of expectation biases: Optimism and overconfidence. Journal
Bethlehem, J., 2010. Selection bias in web surveys. International Statistical Review 78
of Behavioral and Experimental Finance 10, 32–49. https://doi.org/10.1016/j.
(2), 161–188. https://doi.org/10.1111/j.1751-5823.2010.00112.x.
jbef.2016.02.003.
Biais, B., Hilton, D., Pouget, S., Mazurier, K., 2005. Overconfidence, self-monitoring and
Kinari, Y., Tsutsui, Y., 2009. Determinants of Share of Risky Assets in Japan. Review of
trading performance in an experimental financial market. Review of Economic
Monetary and Financial Studies 29, 46–65.
Studies 72, 287–312.
Kitamura, T., Nakajima, K., 2010. Household Portfolio Selection of Businessmen Aged 30
Bianchi, M., 2018. Financial Literacy and Portfolio Dynamics. Journal of Finance 73 (2),
to 49 Years Old (in Japanese). Koudoukeizaigaku (3), 50–69.
831–859. https://doi.org/10.1111/jofi.12605.
Knecht, W.R., 2013. The ”killing zone” revisited: Serial nonlinearities predict general
Black, S.E., Devereux, P.J., Lundborg, P., Majlesi, K., 2018. Learning to Take Risks? the
aviation accident rates from pilot total flight hours. Accident Analysis and
Effect of Education on Risk-Taking in Financial Markets. Review of Finance 22 (3),
Prevention 60, 50–56. https://doi.org/10.1016/j.aap.2013.08.012.
951–975. https://doi.org/10.1093/rof/rfy005.
Kunreuther, H.C., Pauly, M.V., McMorrow, S., 2013. Insurance and behavioral
Brown, M., Grigsby, J., Van Der Klaauw, W., Wen, J., Zafar, B., 2016. Financial
economics: Improving decisions in the most misunderstood industry. Cambridge
Education and the Debt Behavior of the Young. Review of Financial Studies 29 (9),
University Press.
2490–2522. https://doi.org/10.1093/rfs/hhw006.
Liu, E.M., 2013. Time to change what to sow: Risk preferences and technology adoption
Bruhn, M., Leão, L., Legovini, A., Marchetti, R., Zia, B., 2016. The Impact of High School
decisions of cotton farmers in China. Review of Economics and Statistics 95 (4),
Financial Education: Evidence from a Large-Scale Evaluation in Brazil. American
1386–1403. https://doi.org/10.1162/REST_a_00295.
Economic Journal: Applied Economics 8 (4), 256–295.
Lusardi, A., Mitchell, O.S., 2011. Financial Literacy and Planning: Implications for
Burks, S.V., Carpenter, J.P., Goette, L., Rustichini, A., 2009. Cognitive skills affect
Retirement Well-being. Financial literacy: Implications for retirement security and
economic preferences, strategic behavior, and job attachment. PNAS 106 (19),
the financial marketplace. Oxford University Press, New York, pp. 17–39.
7745–7750.
Lusardi, A., Tufano, P., 2015. Debt literacy, financial experiences, and overindebtedness.
Cattaneo, M.D., Crump, R.K., Farrell, M., Feng, Y., 2019. On Binscatter. SSRN Electronic
Journal of Pension Economics & Finance 14 (4), 332–368. https://doi.org/10.1017/
Journal. https://doi.org/10.2139/ssrn.3344739.
s1474747215000232.
Clark, R.L., Morrill, M.S., Allen, S.G., 2012. The role of financial literacy in determining
Mata, R., Frey, R., Richter, D., Schupp, J., Hertwig, R., 2018. Risk Preference: A View
retirement plans. Economic Inquiry 50 (4), 851–866. https://doi.org/10.1111/
from Psychology. Journal of Economic Perspectives 32 (2), 155–172.
j.1465-7295.2011.00390.x.
Mckenzie, C.R.M., Liersch, M.J., Yaniv, I., 2008. Overconfidence in interval estimates :
Deaves, R., Lüders, E., Luo, G.Y., 2009. An experimental test of the impact of
What does expertise buy you ? Organizational Behavior and Human Decision
overconfidence and gender on trading activity. Review of Finance 13 (3), 555–575.
Processes 107, 179–191. https://doi.org/10.1016/j.obhdp.2008.02.007.
https://doi.org/10.1093/rof/rfn023.
Meier, S., Sprenger, C., 2010. Present-Biased and Credit Card Borrowing. American
Dimmock, S.G., Kouwenberg, R., Mitchell, O.S., Peijnenburg, K., 2016. Ambiguity
Economic Journal: Applied Economics 2 (1), 193–210.
aversion and household portfolio choice puzzles: Empirical evidence. Journal of
Meier, S., Sprenger, C.D., 2012. Time discounting predicts creditworthiness.
Financial Economics 119 (3), 559–577. https://doi.org/10.1016/j.
Psychological Science 23 (1), 56–58. https://doi.org/10.1177/0956797611425931.
jfineco.2016.01.003.
Meier, S., Sprenger, C.D., 2013. Discounting financial literacy: Time preferences and
Dohmen, T., Falk, A., Huffman, D., Sunde, U., Schupp, J., Wagner, G.G., 2011. Individual
participation in financial education programs. Journal of Economic Behavior and
risk attitudes: Measurement, determinants, and behavioral consequences. Journal of
Organization 95, 159–174. https://doi.org/10.1016/j.jebo.2012.02.024.
the European Economic Association 9 (3), 522–550. https://doi.org/10.1111/
Melzer, B.T., 2011. The real costs of credit access: Evidence from the payday lending
j.1542-4774.2011.01015.x.
market. Quarterly Journal of Economics 126 (1), 517–555. https://doi.org/10.1093/
Donkers, B., van Soest, A., 1999. Subjective measures of household preferences and
qje/qjq009.
financial decisions. Journal of Economic Psychology 20, 613–642. https://doi.org/
Morse, A., 2011. Payday lenders: Heroes or villains? Journal of Financial Economics 102
10.1016/s1549-3741(04)30005-5.
(1), 28–44. https://doi.org/10.1016/j.jfineco.2011.03.022.
Duflo, E., Saez, E., 2003. The Role of Information and Social Interctions in Retirement
Nogata, D., Takemura, T., 2017. Analysis of risky asset ratios of individual investors
Plan Decisions: Evidence from a Randomized Experiment. Quarterly Journal of
using a web-based survey focusing on the sources of information used by individual
Economics 118 (3), 815–843.
investors [Web ankeito chousa niyoru kojintoshika no kiken shisan hoyuuhiritsu ni
Falk, A., Becker, A., Dohmen, T., Enke, B., Huffman, D., Sunde, U., 2018. Global Evidence
tsuiteno bunseki - kojintoshika no riyou suru jouhougen wo chuushin toshite -] (in
on Economic Preferences. The Quarterly Journal of Economics 133 (4), 1645–1692.
Japanese). Kojinkinyu 17–24.
https://doi.org/10.1093/qje/qjy013.Advance.
Ohtake, F., Akesaka, M., 2017. Personal asset management in Japan and its
Fellner-Röhling, G., Krügel, S., 2014. Judgmental overconfidence and trading activity.
characteristics from a behavioral economics perspective [Nihon no kojinshisan unyo
Journal of Economic Behavior and Organization 107 (PB), 827–842. https://doi.org/
10.1016/j.jebo.2014.04.016.
18
to koudoukeizaigakuteki tokusei] (in Japanese). Shouken Anarisuto Jaanaru (6), Tanaka, T., Camerer, C.F., Nguyen, Q., 2010. Risk and Time Preferences : Linking
16–24. Experimental and Household Survey Data from Vietnam. American Economic
van Rooij, M., Lusardi, A., Alessie, R., 2011. Financial literacy and stock market Review 557–571.
participation. Journal of Financial Economics 101 (2), 449–472. https://doi.org/ The Council for the Promotion of Financial and Economic Education, 2016. Financial
10.1016/j.jfineco.2011.03.006. literacy map [kinyu riterashii mappu] (in japanese).
van Rooij, M.C., Lusardi, A., Alessie, R.J., 2012. Financial Literacy, Retirement Planning The Nikkei, 13 December, 2017. Note the ”true cost” of mutual funds, which can be 4.2
and Household Wealth. Economic Journal 122 (560), 449–478. https://doi.org/ times the trust fee [toushin no hontou no kosuto ni chuui shintakuhoushuu no 4.2
10.1111/j.1468-0297.2012.02501.x. bai mo] (in japanese).https://www.nikkei.com/article/DGXMZO24525310S7A211
Sanchez, C., Dunning, D., 2018. Overconfidence among beginners: Is a little learning a C1NZKP00. Accessed 12 December 2020
dangerous thing? Journal of Personality and Social Psychology 114 (1), 10–28. The, Nikkei, 2020. The reality of installment and revolving loan what you don’t know
Schlösser, T., Dunning, D., Johnson, K.L., Kruger, J., 2013. How unaware are the about credit card loans [bunkatsu ribobarai no genjitsu shiranai to sonsuru
unskilled? Empirical tests of the ”signal extraction” counterexplanation for the kaadoroun] (in japanese).https://style.nikkei.com/article/DGXMZO58080060V10
Dunning-Kruger effect in self-evaluation of performance. Journal of Economic C20A4000000. Accessed 12 December 2020
Psychology 39, 85–100. https://doi.org/10.1016/j.joep.2013.07.004. Watanapongvanich, S., Binnagan, P., Putthinun, P., Khan, M.S.R., Kadoya, Y., 2020.
Sekita, S., 2011. Financial literacy and retirement planning in Japan. Journal of Pension Financial literacy and gambling behavior: Evidence from japan. Journal of Gambling
Economics & Finance 10 (4), 637–656. https://doi.org/10.1017/ Studies 1–21.
S1474747217000270. Yamori, N., Ueyama, H., 2018. Consumer’s financial literacy and financial troubles:
Sekita, S., 2020. Kokumin no shisankeisei to kinyuu riterashii [National asset building Based on the 2016 survey on financial literacy and financial troubles (in japanese).
and financial literacy] (in Japanese). Financial Review 142 (1), 23–41. Journal of Personal Finance and Economics 47, 1–18. https://doi.org/10.18961/
Sekita, S., Kakkar, V., Ogaki, M., 2018. Wealth, Financial Literacy and Behavioral Biases: seikatsukeizaigaku.47.01.
Evidence from Japan. Keio-IES Discussion Paper Series. Yan, J., Kondo, T., Shirasu, Y., Misumi, R., 2019. Nihon no kojintoushika no risukushisan
Shui, H., Ausubel, L. M., 2005. Time Inconsistency in the Credit Card Market. 10.2139/ toushi: kinyuriterashii no shurui ya jouhougen no chigai ha donoyouna eikyou wo
ssrn.586622. ataerunoka? Keieizaimukenkyu 39 (1).
Skimmyhorn, W.L., 2016. Assessing financial education: Evidence from boot camp. Yates, J.F., Ji, L.J., Oka, T., Lee, J.W., Shinotsuka, H., Sieck, W.R., 2010. Indecisiveness
American Economic Journal: Economic Policy 8 (2), 322–343. and culture: Incidence, values, and thoroughness. Journal of Cross-Cultural
Smith, V.L., 1976. Experimental Economics: Induced Value Theory. American Economic Psychology 41 (3), 428–444. https://doi.org/10.1177/0022022109359692.
Review 66, 274–279. https://doi.org/10.1017/cbo9780511528354.008. Yates, J.F., de Oliveira, S., 2016. Culture and decision making. Organizational Behavior
Smith, V.L., 1982. Microeconomic Systems as an Experimental Science. American and Human Decision Processes 136, 1–2. https://doi.org/10.1016/j.
Economic Review 72, 923955. https://doi.org/10.1017/cbo9780511528354.018. obhdp.2016.05.003.
Sutter, M., Kocher, M.G., Glätzle-Rützler, D., Trautmann, S.T., 2013. Impatience and
Uncertainty: Experimental Decisions Predict. American Economic Review 103 (1),
510–531. https://doi.org/10.1257/aer.103.1.510.
19

Journal of The Japanese and International Economies: Tetsuya Kawamura, Tomoharu Mori, Taizo Motonishi, Kazuhito Ogawa

Uploaded by

Document Informationclick to expand document information

Document Informationclick to expand document information

Copyright:

Available Formats

Journal of The Japanese and International Economies: Tetsuya Kawamura, Tomoharu Mori, Taizo Motonishi, Kazuhito Ogawa

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Journal of The Japanese and International Economies: Tetsuya Kawamura, Tomoharu Mori, Taizo Motonishi, Kazuhito Ogawa

Uploaded by

Copyright:

Available Formats

Journal of The Japanese and International Economies 60 (2021) 101131

Contents lists available at ScienceDirect

Journal of The Japanese and International Economies

Is Financial Literacy Dangerous? Financial Literacy, Behavioral Factors, and

1. Introduction techniques and questionnaire methods to reveal individual preferences

(2013) in development economics. Table 2

77 points (10 × X%) or 2 points (100 − 10 × X%) in the risky lottery.

people (who are mostly underconfident) become more unconfident in

the 40s to early 70s on average.

3.4. Econometric framework

Our hypothesis is that financial literacy has both positive and

commonly believed that higher levels of financial literacy induce

appropriate financial behaviors and attitudes, inappropriate ones can

Section 2. Bianchi (2018) points out that financial literacy leads to

holding riskier positions and rebalancing portfolios more frequently. As

negative effects on financial behaviors. Anderson et al. (2017) stress the

importance of perceived, not actual, literacy.

aversion, and ambiguity aversion, we predict that they have negative

effects on all financial behaviors and attitudes, as these factors work to

To test our hypotheses, we estimate the effects of financial literacy,

preferences, and beliefs on financial behaviors and attitudes using the

following regression model:

where Yi is a vector of the financial behaviors and attitudes of respon­

dent i; Ki is a vector of the knowledge and skills, including financial lit­

eracy; Pi is a vector of preference and belief variables; Xi is a vector of

other individual characteristics; α denotes the constant term; and ϵi is an

baseline results in Section 4.1. In Section 4.2, we estimate using the

breakdown of financial literacy test results. In Section 4.3, we conduct

robustness checks on the choice of independent variables using parsi­

choice of estimation methods using ordered probit estimation and IV

Speculative Share of Risky Over- Financial Insufficient Retirement Gambling

Financial Literacy 0.234*** 0.136*** 0.112*** 0.148*** -0.318*** -0.165***

Robust standard errors in parentheses

Speculative Share of Risky Over- Financial Insufficient Retirement Gambling

FL1 Interest 0.051*** 0.000 0.007 0.005 0.003 -0.062***

Robust standard errors in parentheses

Financial Literacy 0.170*** 0.074*** 0.076*** 0.145*** -0.289*** -0.202***

Robust standard errors in parentheses

Speculative Share of Risky Over- Financial Insufficient Retirement Gambling

Financial Literacy 0.236*** 0.159*** 0.099*** 0.144*** -0.335*** -0.166***

Robust standard errors in parentheses

Financial Literacy 0.426*** 0.144*** 0.196*** -0.426*** -0.225***

Robust standard errors in parentheses

Speculative Share of Risky Over- Financial Insufficient Retirement Gambling

Financial Literacy 0.870*** 1.808*** 0.315** 0.530*** -0.943*** -0.535***

Financial Literacy 1.103*** 0.534*** 0.619*** -0.992*** -0.620***

Robust standard errors in parentheses

Fig. 2. Effects of Financial Literacy on Financial Behavior: Binscatter

You might also like

where Yi is a vector of the financial behaviors and attitudes of respon

dent i; Ki is a vector of the knowledge and skills, including financial lit

robustness checks on the choice of independent variables using parsi

Financial Literacy 0.234* 0.136* 0.112* 0.148* -0.318* -0.165*

FL1 Interest 0.051* 0.000 0.007 0.005 0.003 -0.062*

Financial Literacy 0.170* 0.074* 0.076* 0.145* -0.289* -0.202*

Financial Literacy 0.236* 0.159* 0.099* 0.144* -0.335* -0.166*

Financial Literacy 0.426* 0.144* 0.196* -0.426* -0.225***

Financial Literacy 0.870* 1.808* 0.315 0.530* -0.943* -0.535*

Financial Literacy 1.103* 0.534* 0.619* -0.992* -0.620***