ArticlePDF Available


We present results from the first large-scale international survey on time preference, conducted in 53 countries. All countries exhibit hyperbolic discounting patterns, i.e., the immediate future is discounted more than far future. We also observe higher heterogeneity for shorter time horizons, consistent with the pattern reviewed by Frederick, Loewenstein, and O’Donoghue (2002). Cultural factors as captured by the Hofstede cultural dimensions (Hofstede, 1991) contribute significantly to the variation of time discounting, even after controlling for economic factors, such as GDP, inflation rate and growth rate. In particular, higher levels of Uncertainty Avoidance are associated with stronger hyperbolic discounting, whereas higher degrees of Individualism and Long Term Orientation predict stronger tendency to wait for larger payoffs. We also find the waiting tendency is correlated with innovation, environmental protection, crediting rating, and body mass index at country level after controlling for county wealth. These results help us to enhance the understanding of differences across financial markets and economic behavior worldwide.
How Time Preferences Differ:
Evidence from 53 Countries
Mei Wang
Marc Oliver Rieger
Thorsten Hens
April 26, 2016
Published in Journal of Economic Psychology 52 (2016) 115-135
WHU – Otto Beisheim School of Managment, Chair of Behavioral Finance, 56179
Vallendar, Germany,
University of Trier, Chair of Banking and Finance, 54286 Trier, Germany,
University of Zurich, Swiss Finance Institute and Institute of Banking and Finance,
Chair of Financial Economics, Plattenstrasse 32, 8032 Zurich, Switzerland and NHH
Bergen, Norway. Email:
How Time Preferences Differ:
Evidence from 53 Countries
We present results from the first large-scale international survey on
time preference, conducted in 53 countries. All countries exhibit hy-
perbolic discounting patterns, i.e., the immediate future is discounted
more than far future. We also observe higher heterogeneity for shorter
time horizons, consistent with the pattern reviewed by Frederick,
Loewenstein, and O’Donoghue (2002). Cultural factors as captured
by the Hofstede cultural dimensions (Hofstede, 1991) contribute sig-
nificantly to the variation of time discounting, even after controlling
for economic factors, such as GDP, inflation rate and growth rate. In
particular, higher levels of Uncertainty Avoidance are associated with
stronger hyperbolic discounting, whereas higher degrees of Individu-
alism and Long Term Orientation predict stronger tendency to wait
for larger payoffs. We also find the waiting tendency is correlated
with innovation, environmental protection, crediting rating, and body
mass index at country level after controlling for county wealth. These
results help us to enhance the understanding of differences across fi-
nancial markets and economic behavior worldwide.
Keywords: Time preferences; Intertemporal decision; Endogenous prefer-
ence; Cross-cultural comparison.
JEL classification: D90, F40
1 Introduction
Time preference is one of the most fundamental concepts in economics. It
has been widely applied in asset pricing, project evaluation, and decisions on
investment and saving, among many others. Our survey is a first attempt to
collect large-scale empirical data on country-level variations of time prefer-
ences for monetary payoffs. It is to our knowledge the largest international
survey of this kind.
Many factors have been proposed in the literature that could influence
subjective time discounting, such as income, development, culture, and so
forth (Becker & Mulligan, 1997). Given that many of these economic and
cultural factors naturally vary among different countries, it would be very
interesting to test some of the influencing factors in a cross-country sample.
In this article, we elicit time preferences in a large sample across 53 countries
and examine the impacts of culture on time preference.
Studies on cross-cultural differences in temporal discounting are rare.
Most of them involved only two or three countries, e.g., Canadian under-
graduates and foreign undergraduates of Chinese descents (Tan & Johnson,
1996), American, Chinese and Japanese graduate students living in the U.S.
(Du, Green, & Myerson, 2002), and Israeli Arabs and Israeli Jews (Mahajna,
Benzion, Bogaire, & Shavit, 2008).
One problem associated with small samples are confounding factors. Stud-
ies on a limited number of cultural groups have inherent difficulties in dis-
tinguishing the impacts of socio-economic and cultural factors. For example,
the United States and China are different in many dimensions, including
economic situation, political system, and cultural roots. It is hard to de-
duce what causes the observed differences in time preference. To study more
systematically the impacts of country-level factors, it is helpful to include
other countries. For example, including countries like Japan or South Korea,
which have similar cultural roots as China, but a similar economic devel-
opment and political system as the U.S., helps to disentangle these factors.
Including countries in Eastern Europe with different cultural roots, but sim-
ilar modern political experiences as China, is another example how a larger
international sample can provide deeper insights.
The large number of countries included in our survey allows us to link
the measured time preference with the economic and cultural backgrounds of
these countries. We elicit time preferences and time discounting for different
time horizons (one month, one year, and ten years). Our main findings are:
The discount rate for one year is much higher than the discount rate
for ten years: hyperbolic discounting is a global phenomenon.
Time discounting for relatively short time horizons exhibits much higher
heterogeneity than for longer time horizons, consistent with the pattern
noticed by Frederick et al. (2002).
Cultural factors as captured by the Hofstede cultural dimensions (Hofstede,
1991) contribute significantly to the variation of time discounting. In
particular, high levels of Uncertainty Avoidance are associated with
stronger hyperbolic discounting, whereas higher degree of Individual-
ism and Long Term Orientation predict a stronger tendency to wait for
larger payoffs.
We also find that countries with a higher pace of time measured from
field studies (e.g., more punctuality and higher walking speed, as de-
fined by Levine (1997)) are more likely to wait for higher returns, which
provides an external validity for the measurements in our survey.
The collected data on time preferences and time discounting has already
led to many interesting applications, particularly in behavioral finance, such
as applications to the equity risk premium puzzle (Rieger, Wang, & Hens,
2013), dividend payoff policies (Breuer, Hens, Salzmann, & Wang, 2015),
and household debt maturity (Breuer, Rieger, & Soypak, 2014). Institutions
dealing with economic policy issues also find our survey highly valuable. For
example, Marcheggiano and Miles (2013) from the Bank of England used our
data to explain international differences in the effectiveness of fiscal policy.
The rest of this article is organized as follows: In the second section, we
review the literature on culture and time preferences. In the third section,
we present the survey methodology. In the fourth section, we summarize the
key results. In the final section, we discuss possible future research directions
for which this survey data could be used.
2 Relationship between culture and time pref-
Economists traditionally assume preferences are given and there is no role
of culture. As Fehr and Hoff (2011) noted, such views become obsolete with
the growing literature showing that preferences can be endogenous and can
be shaped by societal and cultural influence (Bowles, 1998; Henrich, 2000;
Stern, Dethier, & Rogers, 2005; Eugster, Lalive, Steinhauer, & Zweim¨uller,
2011; Hoff, Kshetramade, & Fehr, 2011).
Perception of time is a part of culture. Culture is typically defined as
something stable over time that distinguishes different groups. Although an
abstract and vague concept to most economists, sociologists and psycholo-
gists have studied in depth the impacts of culture on various aspects, such
as personality, cognition, social and economic development. One of the most
influential measurements for culture has been developed by the Dutch sociol-
ogist Geert Hofstede during his long-term research on cross-national organi-
zational culture. Five persistent cultural dimensions have been found across
different nations and different time periods (Hofstede, 1991, 2001). Here
we discuss three important cultural dimensions related to time preferences,
namely Individualism, Uncertainty avoidance and Long Term Orientation.
Section 3.2 provides more details on the measurement.
Individualism/collectivism is one of the most crucial cultural dimensions
and has been extensively studied. A high score of Individualism implies that
individuals are loosely connected to the society, and are expected to take care
of themselves. In comparison, in a society with collectivistic culture, people
can be protected by some strong cohesive groups throughout lifetime as a
reward to their unshakeable loyalty. The relationship between individualism
and time preference, however, is not clear. On the one hand, the social
connection in a collectivistic culture may provide its citizens a “cushion” or
safety net for potential losses (Hsee & Weber, 1999; Li & Fang, 2004; Weber
& Hsee, 1998), with which people can afford to wait longer and to be more
patient. On the other hand, in an individualistic society, people are expected
to be more independent and to plan their lives by themselves. Triandis
(1971) notes that the “modern man” in a more individualistic culture is more
“concerned with time, planning, willing to defer gratification,” whereas the
“traditional man” in a more collectivistic culture “considers planning a waste
of time, and does not defer gratification.”(p.8.) Therefore, it is also possible
that people in an individualistic culture learn to plan for the future and hence
are more patient. To test the impacts of a collectivistic culture, Mahajna et
al. (2008) compared the subjective discount rates and risk preferences for
Israeli Jews and Arabs with bank customers as participants. Their findings
show that Israeli Arabs, who are supposedly from a more collectivistic society,
have higher subjective discount rates, corresponding to less patience towards
monetary incentives. However, as discussed in the introduction, it is difficult
to disentangle confounding factors with only two cultural groups. Therefore,
with a large sample of countries in our study, we can test more systematically
the relationship between individualism and time preferences, after controlling
other cultural and economic factors.
Uncertainty Avoidance is another cultural dimension relevant to time
preferences. A society with a higher Uncertainty Avoidance score tends to be
less tolerant to uncertain situations. Since future is less predictable than the
present, we expect people from cultures with a higher uncertainty avoidance
tendency to prefer immediate rewards rather than future rewards. To our
best knowledge, no empirical studies have investigated this relationship yet.
The third cultural dimension we study is labeled as “Long Term Ori-
entation.” Hofstede (1991) finds that the Long Term Orientation Score is
typically high in East Asia, especially in Confucian cultures. It implies that
people in such cultures tend to put higher value on the future, and they are
more likely to be patient. Moreover, the concept of “rebirth” in the domi-
nant religions (e.g., Buddhism and Hinduism) in Southeast Asia reflects the
belief that the current life is only a small time interval of one’s entire exis-
tence. Benjamin, Choi, and Strickland (2010) find that priming with Asian
identities makes Asian-American subjects more patient. Chen, Ng, and Rao
(2005) find a similar pattern with bicultural Singaporean participants: par-
ticipants primed with the U.S. culture tend to value immediate consumption
more than Singaporean-primed participants do. However, no previous studies
have directly measured both the Long-term Orientation cultural dimension
and time preferences within the same subject pool as our study does.
3 Methodology
3.1 Measuring time preference
This survey was part of the larger study INTRA (International Test of Risk
Attitudes), conducted by the University of Zurich. The survey contained
three questions on time preferences. The first question was a binary choice
question taken from Frederick (2005), which we refer to as the “wait-or-not”
question in the rest of the article. The question was presented as follows:
Which offer would you prefer?
A. a payment of $3400 this month
B. a payment of $3800 next month
To measure the implicit discount rate more directly, in the next two ques-
tions, we asked participants to give the amount of a delayed payment which
makes them indifferent with an immediate payment. We refer to these two
questions as the “one-year matching question” and the “ten-year matching
question,” respectively. These two questions are1:
1The choice task was chosen from the the first question from Frederick (2005), whereas
the matching task is adapted from two other questions from Frederick (2005). While at
first glance, these numbers seem to be of different orders of magnitude, they are not when
considering the typical answers given by subjects: the median answer for question 3, e.g.,
was $1400 and 25% of the subjects even chose a value of $10,000 or larger. This is of the
same order of magnitude as the amounts in question 1. A starting amount similar to the
Please consider the following alternatives
A. a payment of $100 now
B. a payment of $ Xin one year from now
Xhas to be at least $ , such that B is as attractive as A.
Please consider the following alternatives
A. a payment of $100 now
B. a payment of $ Xin 10 years from now
Xhas to be at least $ , such that B is as attractive as A.
The amount of monetary payoffs in the questions were adjusted accord-
ing to each country’s Purchasing Power Parity (PPP) and the monthly in-
come/expenses of the local students.2
3.2 Measuring cultural dimensions
In the second part of our questionnaire, we used the Values Survey Module
(VSM94) developed by Hofstede and his colleagues to measure the cultural
dimensions (Hofstede, 2001). In particular, we use the results for the follow-
ing three cultural dimensions that are relevant to time discounting:
Individualism (IDV): IDV measures the degree to which the society
reinforces individual or collective achievement, and the extent to which
one-month choice question would therefore have led to much larger amounts.
2The conversion ratio of country ivs. the U.S. is obtained by Xi,U S Pi/PUS , where
Xi,US is the exchange rate of country iand the U.S., Piand PU S are the GDP(PPP) per
capita of country iand the U.S. for the year that the survey was conducted. Addition-
ally, whenever possible, we collected information from difference sources to estimate the
monthly income/expenses of local students (e.g., hourly wage for a student job, typical
food prices in cafeteria, etc.) to double check wether the conversion ratios can be applied
to the university students and in some cases adjusted accordingly.
people are expected to stand up as an individual as compared to loyal
affiliation to a life-long in-group (e.g., extended family, friends, etc.).
The opposite of individualism is collectivism. For example, the U.S. has
an individualistic culture, whereas Japan has a collectivistic culture.
The index is calculated from four questions in our questionnaire where
the participants were asked to rate the importance of the described
feature for an ideal job (1=of utmost importance; 5=of very little or
no importance) : (1) sufficient time for your personal or family life;
(2) good physical working conditions (good ventilation and lighting,
adequate work space, etc.) (3) security of employment; (4) an element
of variety and adventure in the job.
Uncertainty Avoidance (UAI): A high score of UAI indicates that a
society is afraid of uncertain, unknown and unstructured situations. It
is derived from four questions. The first question is “How often do
you feel nervous or tense at work (1=never; 5=always)?” The rest of
the questions asked the participants to what extent they agree with
each of the following statements (1=strongly agree; 5=strongly dis-
agree): (1) One can be a good manager without having precise answers
to most questions that subordinates may raise about their work; (2)
Competition between employees usually does more harm than good;
(3) A company’s or organization’s rules should not be broken – not
even when the employee thinks it is in the company’s best interest.
Long Term Orientation (LTO): When using a Chinese Value Survey
in East Asia, Hofstede (1991) identified a fifth dimension “long-term-
orientation,” or Confucian Dynamism, which captures the society’s
time horizon. It reflects to what extent a society has “a dynamic,
future-oriented mentality.” A higher score implies that the past is val-
ued less than the future, and people may look more forward. We mea-
sure this by asking participants to rate the importance of the following
questions: (1) “In your private life, how important is ‘respect to tra-
dition’ for you (1=of utmost importance; 5=of no importance)?” (2)
“How important is ‘thrift’ for you (1=of utmost importance; 5=of no
There are alternative measures for culture, most notably the Schwartz
dimensions (Schwartz, 2004). They are found to be correlated with Hofstede
dimensions and in order not to stretch the attention of the participants too
much we did not include more than one scale into our questionnaire. There-
fore, we focus on the effects of the Hofstede cultural dimensions which we
measured directly in our survey.
3.3 The survey instrument
A total of 6912 university students in 53 countries/regions participated in
our survey. Most participants were first or second year students from de-
partments of economics, finance and business administration. The average
age of participants was 21.5 years (SD=3.77), and 52.5% of the participants
were males.
Each participant was asked to fill in a questionnaire that included 14
decision making questions (three time preference questions, one ambiguity
aversion question, and 10 lottery questions), 19 questions from the Hofst-
ede VSM94 questionnaire, a happiness question, as well as some information
about their personal background, nationality and cultural origin. The ques-
tionnaire was translated into local languages for each country by professional
translators or translators with economic background. The participants were
instructed that there were no incorrect answers to these questions, and that
the researchers were only interested in their personal preferences and at-
titudes. They were also instructed that they should answer the questions
independently without discussions with others.3In most cases, the survey
was conducted during the first fifteen to twenty minutes of a regular lecture
under the monitoring of the local lecturers and experimenters. The response
rate was therefore very high (nearly 100%) and the number of missing items
relatively small.
After excluding missing responses, the survey yielded 6901 responses for
the first time discounting question, 6608 for the second question, and 6515
for the third question.
3.4 Control variables
Inspired by several studies we decided to include the following control vari-
Becker and Mulligan (1997) proposed a model to capture endogenous time
preferences. It states that the more resources we use to imagine the future,
the more patient we are. It follows that wealth and education leads to pa-
tience. Most studies find wealthier people are more patient (Hausman, 1979;
Lawrance, 1991; Harrison, Lau, & Williams, 2002; Yesuf & Bluffstone, 2008).
Poor farm households, for example, tend to have shorter planning horizons
and hence are reluctant to invest in conservation for natural resources (Mink,
1993). But there are also several studies that find no relation between wealth
and discount rates (Kirby et al., 2002; Anderson, Dietz, Gordon, & Klawit-
ter, 2004). Since we do not have individual wealth or income information,
3The English version of the instruction sheet is available on request.
we use GDP per capita as a proxy for wealth.
Age and gender
A number of experimental and survey studies find that time preferences are
correlated with personal characteristics such as gender (Silverman, 2003) and
age (Green, Fry, & Myerson, 1994). We therefore control for these variables.
Economic growth and inflation
We include the logarithm of the economic growth rate and the annual infla-
tion rate in year 2007, the year before our survey, into the regression analysis.
Since previous times of higher inflation might lead to uncertainty about the
future inflation rate, we repeated all regressions with the log of the maximum
annually inflation rate of the previous ten years, and no significant difference
was found.
4 Results
4.1 Waiting tendency
4.1.1 Descriptive results on waiting tendency
In this section, we evaluate the results from the “wait-or-not” question ($3400
this month or $3800 next month). Table 2 shows the percentage of the partic-
ipants in each country who chose to wait for $3800 next month. We observe
a wide range of variation on the country level – the percentage of students
who chose to wait ranged from only 8% in Nigeria to 89% of Germany. Note
that the implicit interest rate in this question is as high as 11.8% per month
(i.e., an annual discount rate of 280%), which is far higher than the market
interest rates and inflation rates in any of these countries at the time of the
survey. Therefore, the large variation across countries is hard to be justified
Table 1: Overview of countries in the sample
Country N Country N Country N
Angola 57 Germany 540 Norway 192
Argentina 58 Greece 58 Poland 270
Australia 151 Hong Kong 101 Portugal 137
Austria 150 Hungary 262 Romania 339
Azerbaijan 122 India 61 Russia 162
Belgium 46 Ireland 194 Slovenia 96
Bosnia &Herz. 74 Israel 127 South Korea 105
Canada 84 Italy 81 Spain 45
Chile 100 Japan 274 Sweden 65
China 256 Lebanon 101 Switzerland 483
Colombia 147 Lithuania 105 Taiwan 100
Croatia 115 Luxembourg 44 Tanzania 60
Czech Rep 49 Malaysia 99 Thailand 44
Denmark 73 Mexico 89 Turkey 133
Estonia 126 Moldova 100 UK 62
Finland 124 Netherlands 88 USA 72
France 138 New Zealand 91 Vietnam 131
Georgia 38 Nigeria 93 Total 6912
purely by the differences in market interest rates or inflation rates.
In particular, 68% of our U.S. sample chose to wait (N=72). For compari-
son, in the survey by Frederick (2005) where he used the same question with a
relatively large sample (N=807) of U.S. undergraduate students from several
universities, only around 41% of the students chose to wait. Among those stu-
dents who scored high in a separate Cognitive Reflection Test (CRT), there
were 60% choosing the “wait” option, which is closer to our result. The
potential reason is that our participants were studying economics, and thus
Table 2: Percentage of participants choosing the “wait” option
Country Choose to wait Country Choose to wait Country Choose to wait
Germany .89 Lebanon .71 Romania .57
Belgium .87 UK .71 Luxembourg .55
Switzerland .87 Slovenia .71 Moldova .54
Netherlands .85 Ireland .69 Angola .53
Norway .85 Taiwan .69 Vietnam .52
Finland .85 USA .68 Australia .51
Sweden .84 France .65 Azerbaijan .48
Denmark .84 Turkey .64 Spain .47
Czech Rep .80 Argentina .64 Greece .47
Hong Kong .79 China .62 New Zealand .45
Canada .79 Colombia .62 Italy .44
Poland .78 Malaysia .62 Bosnia.Her .39
Austria .78 Portugal .60 Russia .39
Israel .78 Lithuania .60 Chile .37
Estonia .78 India .59 Georgia .26
Hungary .77 Mexico .58 Tanzania .23
Japan .74 Croatia .58 Nigeria .08
South Korea .72 Thailand .57
more likely to take the market interest rate into account. On the other hand,
even 68% of the U.S. sample is still significantly lower than the percentage
in Germanic/Nordic countries like Germany (89%), Switzerland4(87%) or
Finland (86%). This difference is hard to explain only by wealth, education
and the macro-economic situations.5
Figure 1: The percentage of choosing to wait grouped by cultural origin
Note: The column shows the percentage of participants who chose the $3800 option
when they were asked to choose between $3400 this month or $3800 next month.
The respondents were asked about which culture they thought they belong to. We
group the countries into seven cultural clusters inspired by the classification from
Chhokar, Brodbeck, and House (2008).
Each participant has stated not only their nationality, but also the culture
4The survey was conducted in the German language part of Switzerland.
5Even for the students from Princeton University, the percentage choosing the wait
option is lower than the percentage among the German students (80% vs. 89%). Actually
some students from our Norway survey even complained that the question was ridiculous
because “everybody would choose to wait” for one month, given the high implicit interest
they feel they belong to. We classified them into one of seven cultural clus-
ters, mostly following the classification scheme suggested by Chhokar et al.
(2008). Figure 1 shows the percentage of choosing the wait option within each
cultural cluster. In general, the Germanic/Nordic group is far more likely
to wait (85% chose to wait) than other cultural clusters.6Anglo/American,
Middle East, and Asia are similar (around 66% to 68%), followed by East Eu-
rope, then Latin America and Latin Europe. Africa has the lowest percentage
of participants choosing to wait (33%). In the next section, we evaluate to
what extent these differences are related to cultural factors.
4.1.2 Regression results on waiting tendency
We have demonstrated that the responses to such a simple “wait-or-not”
question are highly heterogeneous across countries and across cultural clus-
ters. In the next step, we would like to explore impacts of cultural factors
that correlate with the waiting tendency. To this aim, we control a number
of individual and country-level variables, such as national wealth, gender,
age and so forth.7
We employed mixed-effects multilevel regression with maximum likeli-
hood estimates, with the consideration of interdependence among individuals
within the same country. Table 3 shows the results from multilevel regres-
sions, where the dependent variable is the answer to the waiting question
with country as the group variable. In the remainder of this paper, we use
mixed-effects multilevel regression for other dependent variables as well.
When looking at the impacts of demographic backgrounds, it is interesting
6In fact, the eight countries worldwide with the highest percentage are all from this
cultural cluster – a striking result.
7We also repeated all regression analyses without the non-native students which did
not change the results.
to notice that gender differences play an insignificant role. Although age
turns out to be a significant variable, given the low variation of age among
the student subjects, we consider the variable age only as control and refrain
from making general statements.
On the macroeconomic side, coming from wealthier countries, as mea-
sured by log(GDP/capita), increases the tendency to wait, but other factors
such as growth rate, inflation rate, and economic freedom seem to have little
impact on the waiting tendency.8
Model 2 and 3 in Table 3 indicate the influence of cultural dimensions
after controlling for gender, age and macro-economic variables. Individual-
ism and long-term orientation are robust predictors of the waiting tendency,
both on the country level and on the individual level. The influence of long-
term orientation was as predicted. The effect of individualism is consistent
with the observation by Triandis (1971) in which participants from the more
individualistic culture seemed to be more “willing to defer gratification.” It
is also in line with the findings by Mahajna et al. (2008), where the Israeli
Jews (presumably from a more individualistic culture) exhibited higher pa-
tience for monetary incentives than Israeli Arabs (presumably from a more
collectivistic culture). See Section 2 for more discussion.
Model 3 includes the cultural clusters in the regression. It seems that
even after controlling the Hofstede cultural dimensions, there are still signif-
icant differences across culture clusters. More specifically, participants from
Germanic/Nordic, Anglo/America, Asia, and Middle East cultures are more
willing to wait.
8Since interest rates and GDP per capita from free markets versus state-controlled mar-
kets can be defined very differently, we also repeated all regression analyses by removing
countries with an economic freedom index of less than 60 (mostly unfree and repressed,
according to the official characterization). The results were again basically unchanged.
Table 3: Multilevel Regression on Waiting Tendency
standardized coefficients (t-value)
Independent variables Model 1 Model 2 Model 3
age -0.005*** -0.005*** -0.005***
(-2.70) (-3.01) (-2.96)
gender (male=1) -0.100 -0.014 -0.014
(- 0.90) (-1.25) (-1.26)
inflation rate -0.036 -0.003 -0.002
(-0.48) (-0.51) (-0.42)
Log (growth rate) 0.028 0.051 0.011
(0.69) (1.52) (0.37)
Log (GDP/capita) 0.083*** 0.076*** 0.060***
(3.46) (3.63) (3.08)
Economic freedom 0.001 -0.001 -0.003
(0.24) (-0.24) (-1.45)
Native student dummy 0.055*** 0.053*** 0.017
(2.86) (2.67) (0.79)
Economic major dummy 0.020 0.012 0.009
(1.07) (0.61) (0.46)
IDV average 0.005*** 0.003***
(3.58) (2.65)
IDV ind. diff. 0.000*** 0.000***
(2.65) (2.65)
UAI average -0.001 -0.000
(-1.14) (-0.04)
UAI ind. diff. -0.000 -0.000
(-1.03) (-0.98)
LTO average 0.004* 0.007***
(1.90) (3.26)
LTO ind. diff. 0.001*** 0.001***
(4.48) (4.53)
Africa -0.090
Anglo/America 0.160***
Germ./Nordic 0.172***
L.America 0.019
L.Europe -0.046
E.Europe 0.025
Asia 0.130***
Middle East 0.119**
N 6620 6194 6194
Deviance (-2 log likelihood) 7944.7 7387.4 7347.8
Deviance difference (chi-sqr) 50.73*** 120.75*** 199.27***
Note: 1.* significant at 10% level; **significant at 5% level; ***significant at 1% level; t-values in brackets
2. We denote the country average score of Individualism, Uncertainty Avoidance Index, and Long-term Orientation by
“IDV average,” “UAI average,” and “LTO average.” We denote the difference of individual scores with the country average
score of the respective cultural dimension by “IDV ind. diff.,” “UAI ind. diff.,” and “LTO ind. diff.”
4.2 Inferred Discount Rate: The Classical Approach
To infer discount rates from intertemporal decisions, we use the relationship
between the present value of a cashflow, denoted by P, and its future value,
denoted by F. Formally,
F=P(1 + R)t,
where Ris the discount rate and tis the time to be waited. Since both P
and tare given in our questions, the inferred discount rate can be obtained
easily from
R= (F /P )(1/t)1.
We have two questions (see Section 3.1) to infer the subjective discount rate
(assuming annual compounding), where tequals to 1 year and 10 years, re-
The classical approach states that there is only one “market riskless discount
rate”, which is supposed to be the same for all individuals. Our results in-
dicate that this is not the case. Figure 2 shows the median implicit annual
interest rate for one-year and 10-year matching questions for all countries.9
We observe substantial heterogeneity of the implicit interest rate across indi-
viduals and across countries. The median implicit annual interest rate for the
one year question (R1year) is 100%, ranging from 14% in Australia to 1567%
in Bosnia & Herzegovina, whereas the median implicit annual interest rate
for the ten year question (R10year) is 29%, ranging from 7% in Thailand and
Spain to 73% in Bosnia & Herzegovina.
9We exclude Georgia from this analysis: Georgia had an extremely high implicit interest
rate, especially for the one-year-matching question (14900% for the one-year question, and
86% for the ten-year question). The potential reason is that the survey in Georgia has
been conducted two months before the outbreak of the Russian-Georgian war in 2008.
The feeling of uncertainty induced by the tensions preluding the war may have induced
high discounts for the near future.
For all countries except for Australia, the median R1year is higher than
R10year , which is consistent with the typical empirical findings that discount
rates decrease with longer time horizons. This is also true at the individual
level. In total, 87% participants had an implicit interest rate R1year higher
than R10year . A paired t-test shows that, for all countries except for Finland
and India, the average implicit interest rate for one year is significantly higher
than the implicit interest rate for ten years at 5% level.
The classical discounted utility model assumes consistent time prefer-
ences by using an exponential discounting model. It implies that the time
preference between any adjacent periods should hold constant. Consistent
with previous empirical findings (Thaler, 1981; Benzion, Rapoport, & Yagil,
1989; Du et al., 2002; Frederick et al., 2002), our results show that most peo-
ple in most countries discount the near future more than the far future. This
pattern can be elegantly modelled by the (quasi-)hyperbolic and subadditive
discounting models, which we discuss in more details in the following sections.
4.3 Quasi-hyperbolic and subadditive discounting model
Quasi-hyperbolic discounting model
The quasi-hyperbolic discounting model is usually defined in discrete time
periods as follows:
u(x0, x1, ..., xT) = u(x0) +
This discount function has been used by Phelps and Pollak (1968) to
study intergenerational discounting and by Laibson (1997) to intra-personal
decision problems. When 0 < β < 1 and 0 < δ < 1, people appear to be
more patient in the long run and less patient for the immediate future. The
Figure 2: Median Implicit annual interest rate for 1-year and 10-year horizon
per-period discount rate between now and the next period is (1 βδ)/βδ
and the per-period discount rate between any two future periods is (1 δ),
which is less than (1 βδ)/βδ. The quasi-hyperbolic discounting model as-
sumes a declining discount rate between this period and the next, but a
constant discount rate thereafter. It has often been discussed in the context
of irrationality, such as lack of control, and thus used to justify the need for
commitment devices. In particular, βrefers to the degree of “present bias”.
Larger βimplies less present bias. When β=1, the quasi-hyperbolic discount-
Figure 3: Median values of Parameters in Hyperbolic Discounting Model for
All Countries
ing model coincides with the standard exponential discounting model. We
call the other parameter δthe long-term discount factor.
When we assume a linear utility function, the two matching questions
about time discounting can be represented as:
100 = βδF1year ,
100 = βδ10F10year.
Thus δand βcan be inferred from the responses F1year and F10year:
F10year 1/9
For all participants, the median value of βis 0.60 (Mean=0.56, SD=0.36),
and the median value of δis 0.82 (Mean=0.82, SD=0.12). To reduce the influ-
ence from outlier responses, we have excluded a small number of participants
from the analysis, since their βor δwas large (11 participants with β > 2
and 15 participants with δ > 2), probably by mistake. See Figure 3 for a plot
of parameter estimates of βand δfor each country. Note that the variation
in the present bias discount factor βis much higher than the variation in the
long-term discount factor δ.
Subadditive discounting model
Declining patience can also be explained by subadditive time discount-
ing, i.e., people discount more when the delay is divided into shorter subin-
tervals than when it is undivided (Read, 2001; Read & Roelofsma, 2003;
Scholten & Read, 2010, 2006). Hyperbolic discounting mainly reflects im-
pulsiveness, whereas subadditive discounting mainly reflects perception of
time (Zauberman, Kim, Malkoc, & Bettman, 2009; Read, 2001; Scholten
& Read, 2006). It would imply, in the context of our questions in the sur-
vey, that one year is discounted more than ten years simply because it is a
shorter interval, but not because it is more present. This suggests that time
discounting is not only a function of how far away the outcome is from now,
but also a function of the length of the time interval.
Read (2001) suggests the following subadditive discounting function:
1 + k(TT0)s
where fT0Trepresents the discount factor from time T0to T,kis the hy-
perbolic discounting factor, and sis a parameter that captures the perception
of time.
Again, assuming a linear utility function, the two matching questions
about time discounting can be represented as
100 = F1year
1 + k·1s,100 = F10year
1 + k·10s.
Thus kand scan be inferred from the responses F1year and F10year:
100 1,
s= [log10(log(1+k)(F10year/F1year)] + 1.
4.4 Regression results on quasi-hyperbolic and subad-
ditive discounting factors
Quasi-hyperbolic discounting model
Table 4 and 5 show the results from regression analyses, where the de-
pendent variables are the quasi-hyperbolic time discounting factors βand δ,
respectively. Similar to previous regression results in Table 3, no significant
gender differences are found. Although age is again significant, we don’t
make general conclusions due to the low variation in our sample. GDP per
capita is also a robust predictor and participants from wealthier countries
tend to discount less (higher βand δ) and thus tend to be more patient, con-
sistent with the pattern for the waiting tendency question in Table 3. Growth
rate, inflation rate, and economic freedom seem to have no significant effects,
similar to the findings in Table 3.
Let us now take a look at the cultural variables: for β, we find a strong
effect of Uncertainty Avoidance at both country and individual level (Ta-
ble 4): individuals coming from countries with a high level of UAI and with
higher than country-average UAI tend to have higher present bias. Table 4
also shows that individuals with Long Term Orientation higher than their
country averages are less likely to have present bias, i.e., higher values of β.
Table 4: Multilevel Regression on Present Bias Discount Factor β
standardized coefficients (t-value)
Independent variables Model 1 Model 2 Model 3
age 0.013*** 0.013*** 0.166***
(7.46) (7.13) (7.08)
gender (male=1) 0.013 0.016 0.015
(1.17) (1.40) (1.33)
inflation rate -0.000 -0.000 0.003
(-0.01) (-0.03) (0.27)
Log (growth rate) -0.115 -0.935 -0.122*
(-1.39) (-1.25) (-1.71)
Log (GDP/capita) 0.112** 0.110** 0.119***
(2.36) (2.46) (2.72)
Economic freedom -0.004 -0.007 -0.009*
(-0.70) (-1.42) (-1.94)
Native student dummy 0.017 0.007 0.008
(0.84) (0.32) (0.36)
Economic major dummy 0.051*** 0.432** 0.043**
(2.61) (2.15) (2.16)
IDV average 0.000 0.001
(0.16) (0.19)
IDV ind. diff. 0.000 0.000
(0.15) (0.17)
UAI average -0.007*** -0.006**
(-2.70) (-2.32)
UAI ind. diff. -0.000** -0.000**
(-2.45) (-2.44)
LTO average 0.003 0.004
(0.70) (0.89)
LTO ind. diff. 0.001*** 0.001***
(2.74) (2.76)
Africa -0.821
Anglo/America 0.017
Germanic -0.008
L.America -0.006
L.Europe -0.104
E.Europe -0.101**
Asia 0.048
Middle East -0.021
Number of Obs 6192 5833 5833
Deviance (-2 log likelihood) 7213.5 6744.4 6735.0
Deviance difference (chi-sqr) 82.90*** 107.65*** 122.97***
Note: 1. Larger values imply more patience.
2. * significant at 10% level; **significant at 5% level; ***significant at 1% level; t-values in brackets
3. We denote the country average score of Individualism, Uncertainty Avoidance Index, and Long-term Orientation by
“IDV average,” “UAI average,” and “LTO average.” We denote the difference of individual scores with the country average
score of the respective cultural dimension by “IDV ind. diff.,” “UAI ind. diff.,” and “LTO ind. diff.”
Table 5: Multilevel Regression on Long-term Discount Factor δ
standardized coefficients (t-value)
Independent variables Model 1 Model 2 Model 3
age 0.002*** 0.002*** 0.003***
(3.80) (4.19) (4.24)
gender (male=1) -0.001 0.001 0.001
(-0.35) (0.26) (0.23)
inflation rate -0.001 -0.001 -0.000
(-0.59) (-0.48) (-0.22)
Log (growth rate) 0.003 -0.000 0.001
(0.31) (-0.01) (0.15)
Log (GDP/capita) 0.009* 0.013*** 0.014**
(1.77) (2.64) (2.59)
Economic freedom 0.001 0.000 0.000
(0.91) (0.13) (0.07)
Native student dummy -0.001 -0.001 -0.000
(-0.15) (-0.21) (-0.04)
Economic major dummy 0.004 0.004 0.004
(0.80) (0.79) (0.81)
IDV average -0.000 -0.000
(-1.36) (-1.03)
IDV ind. diff. 0.000 0.000
(0.92) (0.91)
UAI average -0.000 -0.000
(-1.06) (-0.92)
UAI ind. diff. 0.000* 0.000
(1.65) (1.64)
LTO average -0.000 -0.001
(0.57) (-1.15)
LTO ind. diff. 0.000 0.000
(0.38) (0.36)
Africa -0.006
Anglo/America -0.010
Germanic -0.001
L.America 0.010
L.Europe -0.001
E.Europe -0.002
Asia -0.002
Middle East -0.018
Number of Obs 6196 5837 5837
Deviance (-2 log likelihood) 9272.8 8784.0 8787.4
Deviance difference (chi-sqr) 33.23*** 48.04*** 50.56***
Note: 1. Larger values imply more patience.
2. * significant at 10% level; **significant at 5% level; ***significant at 1% level; t-values in brackets
3. We denote the country average score of Individualism, Uncertainty Avoidance Index, and Long-term Orientation by
“IDV average,” “UAI average,” and “LTO average.” We denote the difference of individual scores with the country average
score of the respective cultural dimension by “IDV ind. diff.,” “UAI ind. diff.,” and “LTO ind. diff.”
Table 6: Multilevel regression of hyberbolic discounting rate k
standardized coefficients (t-value)
Independent variables Model 1 Model 2 Model 3
age -0.070*** -0.069*** -0.689***
(-10.29) (-9.98) (-9.95)
gender (male=1) -0.116*** -0.121*** -0.119***
(-2.73) (-2.75) (-2.71)
inflation rate 0.011 0.012 -0.003
(0.22) (0.27) (-0.08)
Log (growth rate) 0.379 0.282 0.377
(1.37) (1.16) (1.61)
Log (GDP/capita) -0.360** -0.357** -0.370**
(-2.25) (-2.43) (-2.54)
Economic freedom 0.001 0.015 0.022
(0.08) (0.92) (1.36)
Native student dummy -0.073 -0.389 -0.288
(-0.96) (-0.50) (-0.33)
Economic major dummy -0.273*** -0.259*** -0.259***
(-3.69) -3.39) (-3.39)
IDV average -0.005 -0.003
(-0.54) (-0.37)
IDV ind. diff. -0.000 -0.000
(-0.22) (-0.22)
UAI average 0.028*** 0.025***
(3.24) (2.92)
UAI ind. diff. 0.001** 0.001**
(2.47) (2.47)
LTO average -0.007 -0.127
(-0.47) (-0.79)
LTO ind. diff. -0.003*** -0.003***
(-2.78) (-2.78)
Africa 0.419
Anglo/America -0.104
Germanic -0.041
L.America 0.161
L.Europe 0.294
E.Europe 0.234
Asia -0.261
Middle East 0.063
Number of Obs 6093 5747 5747
Deviance (-2 log likelihood) 23151.3 21831.0 21822.2
Deviance difference (chi-sqr) 151.38*** 186.70*** 202.27***
Note: 1. Larger values imply more patience.
2. * significant at 10% level; **significant at 5% level; ***significant at 1% level; t-values in brackets
3. The dependent variable is ln(k). We denote the country average score of Individualism, Uncertainty Avoidance Index,
and Long-term Orientation by “IDV average,” “UAI average,” and “LTO average.” We denote the difference of individual
scores with the country average score of the respective cultural dimension by “IDV ind. diff.,” “UAI ind. diff.,” and “LTO
ind. diff.”
Table 7: Multilevel regression of subadditivity factor s
standardized coefficients (t-value)
Independent variables Model 1 Model 2 Model 3
age 0.021*** -0.020*** -0.020***
(-8.62) (-8.27) (-8.29)
gender (male=1) -0.050*** -0.050*** -0.048***
(-3.21) (-3.08) (-2.99)
inflation rate -0.002 -0.001 -0.006
(-0.23) (-0.15) (-0.68)
Log (growth rate) 0.084 0.070 0.078
(1.49) (1.38) (1.63)
Log (GDP/capita) -0.038 -0.032 -0.025
(-1.12) (-1.00) (-0.77)
Economic freedom -0.002 0.000 0.002
(-0.58) (0.05) (0.68)
Native student dummy -0.008 -0.014 -0.010
(-0.29) (-0.48) (-0.31)
Economic major dummy -0.076*** -0.076*** -0.078***
(-2.75) (-2.70) (-2.77)
IDV average -0.001 -0.002
(-0.62) (-0.95)
IDV ind. diff. 0.000 0.000
(1.48) (1.52)
UAI average 0.005*** 0.004**
(2.77) (2.13)
UAI ind. diff. 0.000* 0.000*
(1.80) (1.81)
LTO average -0.000 -0.003
(-0.06) (-0.97)
LTO ind. diff. -0.001*** -0.001**
(-2.62) (-2.59)
Africa 0.166
Anglo/America -0.053
Germanic 0.030
L.America 0.144
L.Europe 0.044
E.Europe 0.169***
Asia -0.029
Middle East 0.077
Number of Obs 5446 5163 5163
Deviance (-2 log likelihood) 9171.8 8723.3 8706.8
Deviance difference (chi-sqr) 106.15*** 131.58*** 157.44***
Note: 1. Larger values imply more patience.
2. * significant at 10% level; **significant at 5% level; ***significant at 1% level; t-values in brackets
3. The dependent variable is ln(k). We denote the country average score of Individualism, Uncertainty Avoidance Index,
and Long-term Orientation by “IDV average,” “UAI average,” and “LTO average.” We denote the difference of individual
scores with the country average score of the respective cultural dimension by “IDV ind. diff.,” “UAI ind. diff.,” and “LTO
ind. diff.”
It is also interesting to see that the long-term discount factor δin Table 5 is in
general not influenced by cultural dimensions, and only affected by country
wealth level (GDP/capita).
Subadditive time discounting model
Table 6 and 7 show the results from regression analyses, where the depen-
dent variables are the hyperbolic time discount rate kand the subadditivity
factor s, respectively. It seems that male and economic students tend to be
more patient (lower k) but have stronger subadditivity (lower s), whereas
older students tend to be more patient and have less subadditivity of time
discounting. Participants from wealthier countries (i.e., higher GDP per
capita) tend to discount less and have lower value of k(Table 6). However,
wealth is not found significant in predicting subadditivity (Table 7). Growth
rate, inflation rate, and economic freedom seem to have no significant effects,
similar to previous results.
Concerning the cultural variables, we find again a strong effect of Uncer-
tainty Avoidance both on the country and on the individual level (Table 6
and Table 7): higher UAI corresponds to higher discounting, but less sub-
additivity. Table 6 also shows that individuals with Long Term Orientation
higher than their country averages discount less, but they may be prone to
stronger subadditivity, as shown in Table 7.
4.5 Partial correlations between different measurement
of time preference
As mentioned before, we are not the first to study the relationship between
culture and time. There is indeed an interesting connection to previous works
in social psychology: Robert Levine defined and measured a concept which
he called “pace of time” in a field study across 31 countries (Levine, 1997).
This overall-pace measure is calculated out of three measures that could be
obtained in most countries: walking speed, postal service speed, and clock
accuracy. Interestingly, we find this measurement is highly correlated with
our measured waiting tendency (ρ= 0.44, p < 0.01) (see Table 8).10
Table 8: Partial correlations between different measurements of time prefer-
ence with log(GDP/cap) as control variable
INTRA Levine Globe
Present bias Long term Time Future orientation
βdiscount δpace (Societal practices)
Waiting tendency 0.23 0.03 0.53** 0.46***
df 50 50 18 32
Present bias β0.49*** 0.01 0.29*
df 50 18 33
Long term discount δ0.03 0.10
df 18 33
1.* significant at 10% level; **significant at 5% level; ***significant at 1% level;
2. Waiting tendency is the percentage of participants in each country who chose to wait one
month in question 1. βand δare the median value of present bias and long-term discount
factor for each country, based on the responses to Question 2 and 3. “Time pace” is
measured by Levine (1997) in his field study to capture the tempo and punctuality in a
country. The lower score implies faster speed and more punctuality. “Future orientation”
is measured by House, Hanges, Javidan, Dorfman, and Gupta (2004), p.304. Higher scores
reflect a more long-term perspective as the accepted norm for the organizations.
This can be most likely understood by considering the discounting effect
for disutilities: an “impatient” person would be very “patiently” procrasti-
10Further regression analyses showed that the time pace measure is significant even when
we control for GDP per capita.
nating some dull or annoying tasks. This attitude would then manifest itself
in slow walking speed, slow and inaccurate service and the tendency to post-
pone tedious tasks like adjusting a watch. We did not have such disutility
questions in our survey, but other surveys found a strong correlation between
impatience for rewards and procrastinating behavior for disutilities (Benzion
et al., 1989). The correlation of Levine’s measurement from the field study
and our measurement from survey questions can be considered as a valuable
cross-validation of both measures.
4.6 Potential applications
In the following, we want to demonstrate the validity and potential usefulness
of our data on four simple examples. Each of them could be taken as a
starting point for further research, based on our survey data.
4.6.1 Innovation
As the first example for possible applications of our measurement, we in-
vestigate whether we can predict a country’s innovation capability by the
measured patience. Technological change and innovation are often treated
as exogenous variables in economic modelling. However, Romer (1990) ar-
gues that it can be endogenously determined. He points out that an increase
in patience will increase research and thus economic growth, which is con-
sistent with the intuition that one must forego some immediate benefits to
invest in research and innovation, in order to get larger rewards in the future.
We test the relationship of patience with the “innovation factor” from the
Global Competitive Report 2008-2009 (Porter & Schwab, 2008). It measures
the technological innovation of a country, in particular investment in research
and development (R & D) in the private sector, the presence of high-quality
scientific research institutions, collaboration in research between universities
and industry, and the protection of intellectual property. We find a positive
correlation between the response of our “wait-or-not” question with the inno-
vation factor at the country level. The first two models in Table 9 show that
after controlling the wealth level of the country, the response to the waiting
question is still highly significant in predicting the innovation factor. This
result suggests that although the wealth level (and hence a general level of a
country’s economy) is crucial to stimulate innovation, the attitude towards
future also plays an important role. For example, while 69% of Taiwanese
participants prefer to wait in the one-month question, only 44% of our Italian
students prefer to wait. The two countries have the same GDP per capita
in 2007, but Taiwan scored much higher in the innovation factor than Italy
(5.26 vs. 4.19). It is worthwhile to investigate further to what extent and
under what mechanism a general attitude towards the future is related to
the innovation activity.
4.6.2 Environmental protection
Studies have revealed that time preference is related to the practice of envi-
ronmental preservation. For example, farmers who discount the future more
strongly were less likely to use soil conservation measures (Yesuf & Bluff-
stone, 2008). Since the wealth level is one important determinant of time
preference, one may argue that we should focus on poverty reduction to make
people discount the future less. However, it is not clear to what extent time
preference per se is a driving factor for a lack of environmental concern. We
illustrate a regression analysis to examine the relative impacts of a country’s
wealth level (as measured by GDP per capita) and the average patience level
(as measured by our first survey question). The dependent variable is the
Table 9: Country-level OLS Regression for Innovation, Gasoline Price, Credit Rating, and Body Mass Index
Dependent Variable
Innovation Factor Gasoline Price Credit Rating Body Mass Index
(1) (2) (3) (4) (5) (6) (7) (8)
Log(GDP/cap) 0.751*** 0.512*** 0.651*** 0.421*** 0.643*** 0.446*** 0.511*** 0.726***
(7.869) (4.435) (5.948) (2.972) (5.872) (3.438) (4.123) (4.660)
Choosing 0.366*** 0.341** 0.328** 0.335**
to wait (3.170) (2.410) (2.530) (2.151)
N49 49 49 49 50 50 49 49
Adjusted R255.4% 62.5% 41.2% 46.6% 40.1% 46.1% 24.6% 29.9%
1.* significant at 10% level; **significant at 5% level; ***significant at 1% level; t-values in brackets
2.The dependent variable “Innovation factor” is from Global Competitive Report 2008-2009 (Porter & Schwab, 2008), pp.18. It measures
the technological innovation of a country, in particular investment in research and development (R&D) in private sectors, the presence of
high-quality scientific research institutions, collaboration in research between universities and industry, and the protection of intellectual
property. “Gasoline price” is from Esty, Levy, Srebotnjak, and Sherbinin (2005), measured by “the ratio of gasoline price to world
average.” “Credit rating” is based on the long-term foreign currency credit rating for sovereign bonds as reported by Standard & Poor’s,
available at of countries by credit rating. “Body Mass Index” is a measure of relative weight based
on an individual’s mass and height, available at mass index#Global statistics.
“Ratio of Gasoline Price to the World Average” from the report of Envi-
ronmental Sustainability Index by Esty et al. (2005). This measure is an
indicator of the degree that environmental externalities have been internal-
ized, and hence reflects the concern on environmental sustainability. Model 3
and 4 in Table 9 demonstrate that our measured time preference has a signif-
icant impact on gasoline prices at the country level, after controling GDP per
capita. Our finding is in line with the experimental study by Hardisty and
Weber (2009), where they find that people discount environmental outcomes
in a similar way to monetary outcomes. This would help policy makers to
understand societal discount rates across countries.
4.6.3 Credit Rating
Empirical evidence shows that individual time preferences are correlated with
credit card borrowing and debt maturity choice (Meier & Sprenger, 2010;
Breuer et al., 2014). In a Diamond-type overlapping generations model,
Buiter (1981) shows that the country with a higher discounting rate runs a
current account deficit. Here we would like to see whether the country-level
time preference measure correlates with the credit rating of sovereign bonds,
which reflects the quality of sovereign bond, the degree of public borrowing,
and the probability of defaults. Model 5 and 6 in Table 9 show that the
country average response to our waiting question is significantly correlated
with the credit rating of sovereign bonds at the country level, again, after
controlling the GDP per capita.
4.6.4 Body Mass Index (BMI)
It is also found that time preference can predict health-related behavior
such as smoking and alcohol consumption, and nutrition intake (Khwaja,
Sloan, & Salm, 2006; Chabris, Laibson, Morris, Schuldt, & Taubinsky, 2008;
Weller, Cook III, Avsar, & Cox, 2008). In particular, Chabris et al. (2008)
and Sutter, Kocher, Gl¨atzle-R¨utzler, and Trautmann (2013) find that time
preference measure elicited from choices in experiments correlates with the
body-mass-index (BMI) for adults and adolescents. Consistent with their
findings, the last two columns in Table 9 shows that the average weighting
tendency can explain a certain degree of cross-country variation for the BMI.
The countries with stronger tendency to wait tend to have a low BMI after
controlling the country wealth level.
Figure 4 provides a graphical overview of the relationship between the
average waiting tendency and the above four variables. The y-axis represents
the residuals of dependent variables after regressing on the GDP per capita.
It indicates that the waiting tendency can explain some variations of the
remaining residuals that can not be explained by the GDP per capita.
4.7 Further Applications
As we mentioned in the introduction, there have already been many appli-
cations of this cross-country comparison, in the field of behavioral finance
where market-level behavior might be influenced by time discounting. Other
applications can be found in economic policy analysis. There are certainly
more questions that could be answered with the help of this data. Here are
two examples:
Buiter (1981) presents a theoretical model using country-level time
preferences to explain the capital movement between countries. The
model has not been tested empirically, but now that would be possible.
Shiller (1999) suggests intergenerational and international risk sharing
in pension system, and emphasizes that the international risk sharing is
rarely discussed. Empirical evidence of the degree of time discounting
across countries can be an important input for such discussions.
All of these examples show that systematic investigations and documen-
tations of time preferences across countries will definitely deepen our under-
standing of the discrepancies across countries, and will also provide policy
makers with useful advice for development at a global level.
5 Discussion
5.1 Interpretation of main results
Our study provides further evidence that hyperbolic discounting is a univer-
sal phenomenon. In general, people are more patient for the distant future
and less patient for the immediate future. Such time inconsistency has also
been found in non-human animals (Du et al., 2002; Fehr, 2002; Green, Fisher,
Perlow, & Sherman, 1981; Mazur, 1987; Rodriguez & Logue, 1988). Studies
from psychological, ecological and neurological perspectives help us to under-
stand the deeper roots of this behavioral pattern (Green & Myerson, 1996;
Camerer, Loewenstein, & Prelec, 2005).
In addition to such general features of time preferences, we have also
documented the systematic variation in time preferences, as compared to the
situational and cultural factors of the countries. As suggested by Shefrin
and Thaler (1988), intertemporal choice can be composed by two opposite
processes – an affective and impulsive process versus a more patient far-
sighted process, which correspond to different parts of the brain, which has
been observed later in various studies on brain reactions (Camerer et al.,
Figure 4: Waiting Tendency vs. Residual of Dependent Variables
0% 0.1% 0.2% 0.3% 0.4% 0.5% 0.6% 0.7% 0.8% 0.9% 1%
0# 0.1# 0.2# 0.3# 0.4# 0.5# 0.6# 0.7# 0.8# 0.9# 1#
0$ 0.1$ 0.2$ 0.3$ 0.4$ 0.5$ 0.6$ 0.7$ 0.8$ 0.9$ 1$
0# 0.1# 0.2# 0.3# 0.4# 0.5# 0.6# 0.7# 0.8# 0.9# 1#
2005). We have essentially three measurements in our survey, with different
time horizons (one month, one year, and ten years).
The waiting question might reflect a more intuitive and impulsive de-
cision process, whereas the long-term discount factor δcorresponds to the
far-sighted decision process. The present bias factor βbased on a short-
term question with a one-year time horizon lies somewhere in the middle
of the spectrum. Consequently, we find different sets of covariates for these
three measures. Whether controlling for macroeconomic conditions or not,
we find strong evidence for a cultural influence on time discounting in the
case of the two “behavioral” measurements (waiting tendency and β). The
effects, however, differ: the waiting tendency is significantly correlated with
the individualism and long term orientation cultural dimensions, whereas
βis more correlated with the uncertainty avoidance. It might be that the
impulsive waiting tendency is more deeply rooted in the life style (individ-
ualistic vs. collectivistic) and the cultural values on the future (long term
orientation). The decision involving a longer time horizon then reflects more
the situational concern rather than the cultural manifestation. Uncertainty
avoidance, as Hofstede (2001) suggested, is more concerned with the situa-
tion, as compared with the other two cultural dimensions, and it seems to
play a more important role on the planning on intermediate terms (e.g., one
year). Our results also show that when it comes to long term decisions (ten
years), people seem to be implicitly more influenced by the macroeconomic
condition such as GDP and growth rate, rather than the cultural factors.
The long-term discount factor shows the least variation among the three
measures, which is more consistent with the traditional economic model and
seems to point to a more “rational” decision process.
According to Graham (1981), the concept of time value of money is rooted
in “linear-separable” views of Anglo-American cultures, who view time as
a continuum stretching from past to present to future. In these cultures,
time is considered to be an essential component of money (e.g., via discount
rate/interest rate), a notion that we know from modern economic and finance
textbooks. Other cultures, however, may have dramatically different views
of time. In particular, Graham (1981) explains that Latin American cultures
perceive time as a circular concept that repeats itself with a cyclical pattern.
This “circular-traditional” view of time is the root of the ma˜nana attitudes
in Mexico and other parts of Latin America, where people’s activities are
much more oriented towards the present than towards the future. Therefore,
immediate rewards are preferred. This may explain the low percentage of sub-
jects who chose to wait in our Latin European and Latin American samples
(Figure 1), even though Latin Europe is closer to Western Europe regarding
economic conditions. Therefore we should be cautious to simply equate the
unwillingness to wait for the larger payoff to a degree of impatience. As
Graham (1981) points out, due to the large difference in the perception of
time, in some cultures, when a person is forced to choose between imme-
diate and future rewards, he may view this not as evaluating alternatives,
because future rewards were perceived as of no real value. “He was essentially
asked if he wanted something or nothing”, and thus, “what one person views
as a choice situation, another views as mandated action.” (Graham, 1981,
p.341) In the one-year and ten-year matching questions, when students were
asked to state the amount of money that makes them indifferent, Latin Eu-
ropean exhibited similar preferences as Germanic/Nordic cultures, whereas
Latin Americans were slightly “less” patient. This again suggests that the
one-month waiting question reflects more a general attitude, whereas the
one-year and ten-year matching questions may be more treated as evaluative
Besides the cultural differences captured by the three aforementioned
Hofstede dimensions, there are of course countless differences that cannot be
captured that easily within a simple survey. We find strong evidence that
these differences also affect time discounting in a significant way: including
dummy variables with cultural clusters into the regression leads to significant
coefficients, especially for the waiting question: Germanic/Nordic subjects,
but also to a lesser extent Asian, Middle East and Anglo-American subjects
showed ceteris paribus more “patience”. These results suggest that beyond
the cultural dimensions by Hofstede, further cultural differences are a key to
the understanding of the heterogeneity in time discounting.
There are other cultural differences that may affect time discounting.
Financial discounting, for instance, is found to be related to a range of psy-
chological variables, such as conscientiousness (Daly, Delaney, & Harmon,
2009). Terracciano et al. (2005) reported that in their sample German
Switzerland, Sweden, Germany, Burkina Faso, and Estonia have the highest
scores on Conscientiousness, whereas Spain, Turkey, Croatia, Chile, and In-
donesia have the lowest scores on Conscientiousness. This again seems to be
consistent with our findings: those countries with higher Conscientiousness
scores are more likely to wait for the delayed larger reward in our one-month
5.2 Methodological Concerns
There could be five major concerns or limitations regarding the survey method
we adopted here. The first is that we only used university students as sub-
jects, not a representative sample of the total population. There are, however,
several advantages of this sample selection: (1) First and second year eco-
nomic students understand better the numeric formulations of lottery and
time-preference questions than the general public, but can still answer the
questions intuitively. (2) Students from economics can also be expected to
play an important role in economics and financial markets in each country
and in the global market. The time and risk preferences we study here are rel-
evant for those finance-related activities. (3) Most importantly, as Hofstede
(1991), a leading researcher in cross-cultural comparisons, emphasized: to
make a cross-national comparison, it is crucial to recruit homogeneous, com-
parable groups from each country in order to control the background variables
as much as possible. University students of economics can be considered as
a sample satisfying these properties well.
The second concern about our survey method might be that we only
elicited hypothetical questions without offering real monetary incentives,
such that participants may not be motivated to give thoughtful answers.
However, researchers who compared directly the real and hypothetical re-
wards did not find clear and systematic differences (Johnson & Bickel, 2002;
Coller & Williams, 1999; Kirby & Marakovic, 1995). Moreover, hypothetical
questions have even some advantages in the domain of time preferences be-
cause they allow to ask questions involving a long time span and large payoffs
(Frederick et al., 2002).
A third concern is whether the sample size in each country is sufficiently
large and representative. Previous research shows that even within the same
country, the cultural difference can be very large, e.g., Talhelm et al. (2014).
Indeed, we would have been happy to have large sample sizes in every country,
but in a few countries this was infeasible. However, the total number of
subjects is substantial and the large number of countries from which we
collected data allows to test competing factors on the country level that a
study in fewer countries with a larger subject pool in each country could
not achieve. Moreover, we have shown that the between-country variation,
as compared to the within-country variation, is large enough to justify our
A further concern is whether time preferences can be elicited indepen-
dently from the interest rates of the markets to which the respondents have
access to. One can argue that in a perfect capital market where individu-
als can borrow and lend freely, the personal taste concerning time preference
cannot be elicited, because intertemporal choices are made such that the per-
sonal discount rate corresponds to the interest rate in the market. If markets
were perfect and people answered the question relating the stated monetary
amounts to the borrowing and lending opportunities in these markets then
we would find the discount rates measured in our survey to equal market
Many studies, however, have shown that stated discount rates tend to
be much larger (compare the survey of Frederick (2005)). One of the rea-
sons might be that in reality markets are far from perfect: even in countries
with well-developed financial systems there are many constraints, particu-
larly on borrowing money. They can be institutional or cultural in nature: in
some countries, obtaining a loan might be impossible for many people (com-
pare Beck, Demirg¨c-Kunt, and Peria (2008) for an international comparison
study on this issue), whereas in other countries taking a loan for consumption
might be considered simply as foolish behavior that could reduce reputation
substantially. Another reason might be that the respondents understood the
difference between their personal time preference and the market interest
rate and answered the question applying the former, i.e. without considering
the borrowing and lending opportunities offered in the capital market they
have access to.
A final limitation of our study is that it only focuses on time preferences
in gains, and it is not clear whether the results can be generalised to in-
tertemporal choices involving losses. Normative economic theories prescribe
people to discount both future gains and losses due to opportunity cost and
uncertainty, and there should be no differences in discount rates in gains and
losses. However, a general finding is that gains are discounted more than
losses. This has been called the sign effect (Benzion et al., 1989; Thaler,
1981; Yates & Watts, 1975). More strikingly, a substantial proportion of
participants even prefer a sooner loss to a later loss of the same or smaller
size, showing none or even a negative discounting tendency, e.g., Hardisty and
Weber (2009), Hardisty, Appelt, and Weber (2013), Loewenstein (1987), Sun
et al. (2015), van der Pol and Cairns (2000).
The underlying determinants of time discounting, including neural, psy-
chological, social and economic factors, can be very different in gains and
losses, and the interaction is complicated. Xu, Liang, Wang, Li, and Jiang
(2009) demonstrate that although intertemporal choices in gains and losses
both activate brain regions that are related to high-level cognitive processes,
discounting future losses leads to greater activation of brain regions related
to negative emotions. This is in line with the finding by Hardisty and We-
ber (2009): higher CRT (cognitive reflection test) scores are related to less
discounting in gains but have no effects in losses.
Both studies seem to imply that discounting losses is a more affective than
cognitive process. Moreover, Hardisty and Weber (2009) suggest that social
and cultural norms typically encourage people to wait for larger later gains,
but to avoid larger later losses (patient in both gains and losses), whereas
the fixed-cost present bias as documented by Benhabib, Bisin, and Schot-
ter (2010) (i.e., psychological desire to resolve events immediately) may lead
people to accept both immediate gains and losses (impatient in gains but
patient in losses). Therefore, we expect that cultural factors such as Long-
Term Orientation should increase patience in both gains and losses (wait
for later gains and avoid larger later losses), whereas factors that are more
related to psychological dread such as Uncertainty Avoidance may cause peo-
ple to want immediate gains and losses (i.e., decreases patience in gains but
increases patience in losses). Moreover, a recent neutral study by Tanaka,
Yamada, Yoneda, and Ohtake (2014) report that participants with the sign
effect exhibit stronger brain activity to magnitude and delay of losses than
that of gain, suggesting loss aversion as potential mechanism to sign effect.
Using INTRA survey, Wang, Rieger, and Hens (2016) report the relation be-
tween culture and loss aversion, and discuss the potential role of emotional
regulation. Taken together, we may also deduce that countries with stronger
loss aversion also tend to exhibit stronger sign effect due to the variation
of emotion regulation shaped via cultural influence. We encourage future
research to establish further hypotheses and test empirically the relation of
culture and time discounting in both gains and losses to help us understand
the social and psychological mechanisms for time discounting.
6 Conclusion
We report an international survey on time preference across 53 countries. Our
results are consistent with the previous hyperbolic discounting literature in
that all countries exhibit stronger discounting for one year than for ten years.
More importantly, there is a large cross-country variation in responses to time
preference questions, especially to questions concerning waiting tendency and
short-term discounting rate. Several Hofstede cultural dimensions are cor-
related to the time preference measure. We suggest several applications by
using time preference to explain cross-country difference innovation, environ-
mental protection, credit rating, and health-related behavior as reflected in
the BMI.
Several independent variables in our regression models were endogenous.
Ideally, the parameters should have been estimated by using a simultaneous
equation system. With our cross-section data, it is difficult to identify in-
strumental or lagged variables for such analysis. If time series data could
be collected in the future, then one might gain more insights about causal
relationships. Despite these limitations, this study sheds light on several
important aspects of time preference.
We thank Herbert Dawid, Erich Gundlach, Volker Kr¨atschmer, Rolf J. Langham-
mer, and Daniel Schunk for their comments. We thank Julia Buge, Chun-Houh
Chen, Shiyi Chen, Mihnea Constantinescu, Simona Diaconu, Oliver Dragicevic,
Anke Gerber, Wolfgang H¨ardle, Ljilja Jevtic, Renata Kovalevskaja, Dana Liebman,
Takeshi Momi, Andres Mora, Koji Okada, Hersh Shefrin, Fangfang Tang, Bodo
Vogt, Hannelore Weck-Hannemann, T˜onn Talpsepp, Evgeny Plaksen, Longinus
Rutasitara, Xiao-Fei Xie, Levon Mikayelyan, Andres Mora, Ante Busic, Alexan-
der Meskhi, Christos Iossifidis, Janos Mayer, Istvan Laszloffy, Stephan Passon,
Salim Cahine, Renata Kovalevskaja, Besart Colaku, Simona Diaconu, Thierry
Post, Bjørn Sandvik, Ermira Mehmetaj, Aleksandra Przywuska, Simona Diaconu,
Sonja Ratej Pirkovic, Antonio Avillar, Rosemarie Nagel, Pattarake Sarajoti, Haluk
Bilge Halas, Markus K. Brunnermaier, Jing Qian, Markus Leippold, Thuy Bui,
Surajit Sinha and numerous other people for generous help on data collection and
We would like to thank the following universities for participation in the sur-
vey: Catholic University of Angola (Angola), Universidad Torcuato Di Tella (Ar-
gentina), Universit¨at Innsbruck (Austria), Alpen-Adria-Universit¨at Klagenfurt (Aus-
tria), University of Adelaide (Australia), Khazar University (Azerbaijan), Catholic
University in Leuwen (Belgium), Pan-European University Apeiron (Bosnia and
Herzegovina), University of Windsor (Canada), University of British Columbia
(Canada), Fudan University (China), Peking University (China), Renmin Univer-
sity (China), Universidad de Chile, Universidad de los Andes (Colombia), Buise-
ness College Vern’ (Croatia), CERGE-EI (Czech Rep.), University of Southern
Denmark, University of Copenhagen (Denmark), Tallinn University of Technology
(Estonia), University of Vaasa (Finland), Universit¨at Hamburg (Germany), Uni-
versit¨at Trier (Germany), Universit¨at Konstanz (Germany), Otto-von-Guericke
Universit¨at Magdeburg (Germany), University of Thessaly (Greece), Hong Kong
Chinese University, Hong Kong Baptist University (Hong Kong), University of
ecs (Hungary), Indian Institute of Technology Kanpur (India), Ben Gurion Uni-
versity (Israel), NUI Maynooth (Ireland), Universit`a degli Studi di Venezia (Italy),
Foreign Trade University (Vietnam), Doshisha University (Japan), American Uni-
versity of Beirut (Lebanon), Vilnius University (Lithuania), University of Malaya
(Malaysia), Universidad de Guanajuato (Mexico), MAES Kishinev (Moldova),
Maastricht University (Netherlands), Massey University (New Zealand), Univer-
sity of Ibadan (Nigeria), NHH Bergen (Norway), Warsaw University of Technology
and Engineering (Poland), University of Lisboa (Portugal), Bucharest Academy
of Economic Studies (Romania), Russian Customs Academy Vladivostok (Rus-
sia), University of Ljubljana (Slovenia), Seoul National University (South Korea),
Universidad Pablo de Olavide (Spain), University of Zurich (Switzerland), Na-
tional Sun Yat-sen University (Taiwan), University of Dares Salaam (Tanzania),
Chulalongkorn University (Thailand), Middle East Technical University (Turkey),
Bogazici University (Turkey), Keele University (UK), Emory University (USA),
Santa Clara University (USA), Princeton University (USA).
Financial support by the National Centre of Competence in Research “Financial
Valuation and Risk Management” (NCCR FINRISK), Project 3, “Evolution and
Foundations of Financial Markets”, and by the University Research Priority Pro-
gram “Finance and Financial Markets” of the University of Z¨urich is gratefully
Anderson, C. L., Dietz, M., Gordon, A., & Klawitter, M. (2004). Discount
rates in Vietnam. Economic Development and Cultural Change,52 (4),
Beck, T., Demirg¨c-Kunt, A., & Peria, M. S. M. (2008, November 7). Bank-
ing services for everyone? Barriers to bank access and use around the
world. The World Bank Economic Review,22 (3), 397-430.
Becker, G. S., & Mulligan, C. B. (1997, Aug.). The endogenous determination
of time preference. The Quarterly Journal of Economics,112 (3), 729-
Benhabib, J., Bisin, A., & Schotter, A. (2010). Present-bias, quasi-hyperbolic
discounting, and fixed costs. Games and Economic Behavior,69 (2),
Benjamin, D. J., Choi, J. J., & Strickland, A. J. (2010). Social identity and
preferences. Amercian Economic Review ,100 , 1913-1928.
Benzion, U., Rapoport, A., & Yagil, J. (1989, Mar.). Discount rates inferred
from decisions: An experimental study. Management Science,35 (3),
Bowles, S. (1998). Endogenous preferences: The cultural consequences of
markets and other economic institutions. Journal of Economic Litera-
ture,36 (1), 75-111.
Breuer, W., Hens, T., Salzmann, A. J., & Wang, M. (2015). On the deter-
minants of household debt maturity choice. Applied Economics,47 (5),
Breuer, W., Rieger, M. O., & Soypak, C. (2014). The behavioral founda-
tions of corporate dividend policy: A cross-country empirical analysis.
Journal of Banking and Finance,42 , 247-265.
Buiter, W. H. (1981). Time preference and international lending and bor-
rowing in an overlapping-generations model. The Journal of Political
Economy,89 (4), 769-797.
Camerer, C., Loewenstein, G., & Prelec, D. (2005). Neuroeconomics: How
neuroscience can inform economics. Journal of Economic Literature,
43 (1), 9-64.
Chabris, C. F., Laibson, D., Morris, C. L., Schuldt, J. P., & Taubinsky,
D. (2008). Individual laboratory-measured discount rates predict field
behavior. Journal of Risk and Uncertainty,37 (2-3), 237-69.
Chen, H., Ng, S., & Rao, A. R. (2005). Cultural differences in consumer
impatience. Journal of Marketing Research,42 , 291-301.
Chhokar, J. S., Brodbeck, F. C., & House, R. J. (2008). Culture and lead-
ership across the world: The GLOBE book of in-depth studies of 25
societies. New York: Taylor & Francis Group.
Coller, M., & Williams, M. (1999). Eliciting individual discount rates. Ex-
perimental Economics,2(107-127).
Daly, M., Delaney, L., & Harmon, C. P. (2009). Psychological and biological
foundations of time preference. Journal of the European Economic
Association,7(2-3), 659-669.
Du, W., Green, L., & Myerson, J. (2002). Cross-cultural comparisons on dis-
counting delayed and probabilistic rewards. The Psychological Record,
52 , 479-492.
Esty, D., Levy, M., Srebotnjak, T., & Sherbinin, A. de. (2005). 2005 envi-
ronmental sustainability index: Benchmarking national environmental
stewardship. New Haven: Yale Center for Environmental Law & Policy.
Eugster, B., Lalive, R., Steinhauer, A., & Zweim¨uller, J. (2011). The demand
for social insurance: does culture matter? Economic Journal,121 ,
Fehr, E. (2002). The economics of impatience. Nature,415 (17), 269-272.
Fehr, E., & Hoff, K. (2011). Introduction: Tastes, castes and culture: The
influence of society on preferences. Economic Journal,121 , 396-412.
Frederick, S. (2005). Cognitive reflection and decision making. Journal of
Economic Perspectives,19 (4), 25-42.
Frederick, S., Loewenstein, G., & O’Donoghue, T. (2002). Time discounting
and time preference: a critical review. Journal of Economic Literature,
40 , 350-401.
Graham, R. J. (1981). The role of perception of time in consumer research.
Journal of Consumer Research,7, 335-342.
Green, L., Fisher, E. B., Perlow, S., & Sherman, L. (1981). Preference
reversal and self-control: Choice as a function of reward amount and
delay. Behaviour Analysis Letters,1(43-51).
Green, L., Fry, A. F., & Myerson, J. (1994). Discounting of delayed rewards:
A life-span comparison. Psychological Science,1, 33-36.
Green, L., & Myerson, J. (1996). Exponential versus hyperbolic discounting
of delayed outcomes: risk and waiting time. Ameri. Zool.,36 , 496-505.
Hardisty, D. J., Appelt, K. C., & Weber, E. U. (2013). Good or bad, we want
it now: Fixed-cost present bias for gains and losses explains magnitude
asymmetries in intertemporal choice. Journal of Behavioral Decision
Making,26 , 348-361.
Hardisty, D. J., & Weber, E. U. (2009). Discounting the future green: Money
versus the environment. Journal of Experimental Psychology: General,
138 (3), 329-340.
Harrison, G. W., Lau, M. I., & Williams, M. B. (2002). Estimating individual
discount rates in Denmark: A field experiment. American Economic
Review,92 (5), 1606-1617.
Hausman, J. A. (1979). Individual discount rates and the purchase and
utilization of energy-using durables. The Bell Journal of Economics,
10 (1), 33-54.
Henrich, J. (2000). Does culture matter in economic behavior? Ultimatum
game bargaining among the Machiguenga of the Peruvian Amazon.
American Economic Review ,90 (4), 973-979.
Hoff, K., Kshetramade, M., & Fehr, E. (2011). Caste and punishment:
the legacy of caste culture in norm enforcement? Economic Journal ,
121 (556), F449-F475.
Hofstede, G. (1991). The confucius connection: From cultural roots to
economic growth. Organization Dynamics,16 (4), 4-18.
Hofstede, G. (2001). Culture’s consequences, comparing values, behaviors,
institutions, and organizations across nations. Thousand Oaks CA:
Sage Publications.
House, R. J., Hanges, P. J., Javidan, M., Dorfman, P. W., & Gupta, V.
(Eds.). (2004). Culture, leadership, and organizations: The GLOBE
study of 62 societies. Thousand Oaks CA: Sage Publications.
Hsee, C. K. H. K., & Weber, E. U. (1999). Cross-national differences in
risk preference and lay predictions. Journal of Behavioral Deicision
Making,12 , 165-179.
Johnson, M. W., & Bickel, W. K. (2002). Within-subject comparison of real
and hyperthetical money rewards in delay discounting. Journal of the
Experimental Analysis of Behavior ,77 , 129-146.
Khwaja, A., Sloan, F., & Salm, M. (2006). Evidence on preferences and
subjective beliefs of risk takers: The case of smokers. International
Journal of Industrial Organization,24 , 667-682.
Kirby, K. N., Godoy, R., Reyes-Garcia, V., Byron, E., Apaza, L., Leonard,
W., et al. (2002). Correlates of delay-discount rates: Evidence from
Tsimane Amerindians of the Bolivian rain forest. Journal of Economic
Psychology,23 , 291-316.
Kirby, K. N., & Marakovic, N. N. (1995). Modeling myopic decisions: Ev-
idence for hyperbolic delay discounting with subjects and amounts.
Organizational Behavior and Human Decision Processes,64 , 22-30.
Laibson, D. (1997). Golden eggs and hyperbolic discounting. Quarterly
Journal of Economics,112 (2), 443-477.
Lawrance, E. C. (1991, Feb.). Poverty and the rate of time preferences:
Evidence from panel data. The Journal of Political Economy,99 (1),
Levine, R. V. (1997). A geography of time: The temporal misadventures of
a social psychologist, or how every culture keeps time just a little bit
differently. Basic Books, A Member of the Perseus Books Group.
Li, S., & Fang, Y. (2004). Respondents in Asian cultures (e. g., Chinese) are
more risk-seeking and more overconfident than respondents in other
cultures (e. g., in United States) but the reciprocal predictions are in
total opposition: How and why? Journal of Cognition and Culture,
4(2), 263-292.
Loewenstein, G. (1987). Anticipation and the valuation of delayed consump-
tion. Economic Journal,97 (387), 666-684.
Mahajna, A., Benzion, U., Bogaire, R., & Shavit, T. (2008). Subjective
discount rates among Israeli Arabs and Israeli Jews. The Journal of
Socio-Economics,37 (6), 2513-2522.
Marcheggiano, G., & Miles, D. (2013). Fiscal multipliers and time preference
(Discussion Paper No. 39). Bank of England.
Mazur, J. E. (1987). An adjusting procedure for studying delayed reinforce-
ment. In M. L. Commons, J. E. Mazur, J. A. Nevin, & H. Rachlin
(Eds.), Quantitative analyses of behavior: Vol. 5. the effect of delay
and of intervening events on reinforcement value (p. 55-73). Hillsdale,
NJ: Erlbaum.
Meier, S., & Sprenger, C. (2010). Present-biased preferences and credit card
borrowing. American Economic Journal: Applied Economics,2(1),
Mink, S. (1993). Poverty and the environment. Finance and Development,
30 (4), 8-10.
Phelps, E. S., & Pollak, R. A. (1968). On second-best national saving and
game-equilibrium growth. Review of Economic Studies,35 , 185-199.
Porter, M. E., & Schwab, K. (2008). The global competitiveness report 2008-
2009 (Tech. Rep.). Geneva, Switzerland: World Economic Forum.
Read, D. (2001). Is time-discounting hyperbolic or subadditive? Journal of
Risk and Uncertainty,23 (1), 5-32.
Read, D., & Roelofsma, P. H. M. P. (2003). Subadditive versus hyperbolic
discounting: A comparison of choice and matching. Organizational
Behavior and Human Decision Processes,91 (2), 140-153.
Rieger, M. O., Wang, M., & Hens, T. (2013). International evidence on the
equity premium puzzle and time discounting. Multinational Finance
Journal,17 (3/4), 1-15.
Rodriguez, M. L., & Logue, A. W. (1988). Adjusting delay to reinforcement:
Comparing choice in pigeons and humans. Journal of Experimental
Psychology: Animal Behavior Processes,14 (1), 105-117.
Romer, P. M. (1990). Endogenous technological change. Journal of Political
Economy,98 (2), S71-S102.
Scholten, M., & Read, D. (2006). Discounting by intervals: A generalized
model of intertemporal choice. Management Science,52 (9), 1424-1436.
Scholten, M., & Read, D. (2010). The psychology of intertemporal tradeoffs.
Psychological Review,117 (3), 925-944.
Schwartz, S. H. (2004). Mapping and interpreting cultural differences around
the world. In H. Vinken, J. Soeters, & P. Ester (Eds.), Comparing
cultures, dimensions of culture in a comparative perspective. Leiden,
The Netherlands: Brill.
Shefrin, H., & Thaler, R. H. (1988). The behavioral life-cycle hypothesis.
Economic Inquiry,26 (4), 609-643.
Shiller, R. (1999). Social security and institutions for intergenerational,
intragenerational, and international risk-sharing. Carnegie-Rochester
Conference Series on Public Policy,50 , 165-204.
Silverman, I. W. (2003). Gender differences in delay of gratification: A meta
analysis. Sex Roles,49 (9/10), 451-463.
Stern, N., Dethier, J.-J., & Rogers, F. H. (2005). Growth and empowerment:
Making development happen. Cambridge, MA: MIT Press.
Sun, H.-Y., Li, A.-M., Chen, S., Zhao, D., Liang, Z.-Y., & Li, S. (2015).
Pain now or later: An outgrowth account of pain-minimization. PloS
one,10 (3), e0119320.
Sutter, M., Kocher, M. G., Gl¨atzle-R¨utzler, D., & Trautmann, S. T. (2013).
Impatience and uncertainty: Experimental decisions predict adoles-
cents’ field behavior. Amercian Economic Review,103 (1), 510-531.
Talhelm, T., Zhang, X., Oishi, S., Shimin, C., Duan, D., Lan, X., et al.
(2014). Large-scale psychological differences within China explained
by rice versus wheat agriculture. Science,344 (6184), 603-608.
Tan, C. T., & Johnson, R. D. (1996). To wait or not to wait: The influence
of culture on discounting behavior. In W. H. Loke (Ed.), Perspectives
on judgment and decision making. Scarecrow: Lanham, MD.
Tanaka, S. C., Yamada, K., Yoneda, H., & Ohtake, F. (2014). Neural mech-
anisms of gain–loss asymmetry in temporal discounting. The Journal
of Neuroscience,34 (6), 5595-5602.
Terracciano et al., A. (2005). National character does not reflect mean
personality trait level in 49 cultures. Science,310 , 96-100.
Thaler, R. H. (1981). Some empirical evidence on dynamic inconsistency.
Economic Letter,8, 201-207.
Triandis, H. C. (1971). Some psychological dimensions of modernization
(Paper presented at the 17th Congress of Applied Psychology). Liege,
van der Pol, M. M., & Cairns, J. A. (2000). Negative and zero time preference
for health. Health Economics,9(2), 171-175.
Wang, M., Rieger, M. O., & Hens, T. (2016). The impact of culture on loss
aversion. Journal of Behavioral Decision Making,forthcoming.
Weber, E. U., & Hsee, C. (1998). Cross-cultural differences in risk percep-
tion, but cross-cultural similiarities in attitudes towards perceived risk.
Management Science,44 (9), 1205-1217.
Weller, R. E., Cook III, E. W., Avsar, K. B., & Cox, J. E. (2008). Obese
women show greater delay discounting than healthy-weight women. Ap-
petite,51 (3), 563-569.
Xu, L., Liang, Z.-Y., Wang, K., Li, S., & Jiang, T. (2009). Neural mechanism
of intertemporal choice: From discounting future gains to future losses.
Brain Research,1261 , 65-74.
Yates, F. J., & Watts, R. A. (1975). Preferences for deferred losses. Organi-
zational Behavior and Human Performance,13 (2), 294-306.
Yesuf, M., & Bluffstone, R. (2008). Wealth and time preference in ru-
ral Ethiopia (EfD discussion paper No. 08-16). Washington DC: the
Environment for Development Initiative and Resources for the Future
Zauberman, G., Kim, B. K., Malkoc, S. A., & Bettman, J. R. (2009). Dis-
counting time and time discounting: Subjective time perception and
intertemporal preferences. Journal of Marketing Research,46 , 543-
... On the one hand, social connection in a collectivist culture may provide its citizens with a "cushion" or safety net in the event of a loss. [12]. Bellman et al.(2004 ) developed an alternative theory implying that such societies will be less concerned with information privacy and will feel less compelled to seek government intervention [13]. ...
We experimentally investigate the link between individuals' value on future incomes and their support for redistributive policies today. The investigation identifies time preferences as a key driver of redistributive policies via their effect on personal responsibility, defined as costly but productive effort. The investigation also accounts for the strategic interplay between individuals in the choice of effort, identifying two key strategies: either exercise more effort and ask for less redistribution or free ride on others’ efforts, asking for more redistribution. We find that individuals oriented toward the future tend to invest more and ask for less redistribution. We discuss the policy implications of this result.
Employing an international sample of banks spanning 2009 to 2018, this paper investigates the effect of individualism on bank risk. We find that individualism is negatively associated with bank risk. We then find that managers’ (individual) risk preference as internal corporate governance and bank regulation as external corporate governance exhibit a mediation effect in explaining the relationship between individualism and bank risk. We further test the effect of individualism on bank performance, including bank efficiency and profitability. Our baseline results remain valid after carefully considering various robustness tests and endogeneity concerns.
Low insurance take‐up in low‐income populations is not easily explained by the standard single‐period expected utility model of insurance that overlooks the relevance of time preference when liquidity is constrained. We design field survey instruments to elicit quasi‐hyperbolic time preferences, as well as prospect theory risk preferences, and use them to examine whether time preferences explain health insurance behavior of low‐income Filipinos. Consistent with theory, those with stronger parameterized time preference are less likely to insure and the partial association is most pronounced at low wealth where liquidity is most likely to be constrained. Among those with better understanding of insurance, lower take‐up is also associated with present bias. We do not find that insurance is significantly associated with risk preferences.
Mankind's ability to mitigate and adapt to climate change may be limited by cognitive biases. To address this challenge, research on cognitive biases has to be expanded beyond the study of individual-based psychological cognition effects to understand their interaction with cultural factors and their impact on group behaviour. Here we describe the relevant cognitive biases and how they are impacted by culture, and we propose that future environmental policymaking has to take into account how such Culturally Embedded Cognitive Biases (CECB) affect willingness to comply.
Introduction: Cultural orientation and interdependent self-construal can moderate the relationship between perceived pro-generation investment and future orientation of young adults. To test how interdependent self-construal moderate the relationship between pro-generation investment and future orientation of young adults from two different cultural ecologies was the aim of the current study. Methods: A cross-cultural comparison was conducted among study participants from China, Germany, and the United States. Interdependent self-construal, perceived pro-generation investment by parents (parental investment), and future orientation were measured. Cross-cultural data were collected from 205 college students in China, a collectivist culture, and 169 college students in Germany (n = 50) and the United States (n = 119), which are individualist cultures. We examined a three-way interaction with cultural orientation and interdependent self-construal as moderators in the relationship between perceived parental investment and future orientation. Results: In the collectivist cultural context, there appeared no moderating effect of interdependent self-construal on the relationship between perceived parental investment and future orientation, although interdependent self-construal and perceived parental investment predicted future orientation. In the individualistic cultural context, there was a moderating effect. For individuals high in interdependent self-construal, future orientation remained stable as perceived parental investment increased. For individuals low in interdependent self-construal, future orientation decreased as perceived parental investment increased. Conclusions: The findings have practical implications in that parents should follow the cultural orientation of their background and provide their children with individualized investment and education to shape the future orientation of their offspring.
Dieser Beitrag schließt sich unmittelbar an den Beitrag von Reinhard H. Schmidt an und beleuchtet die Entwicklungen im Bereich Finance der jüngeren Vergangenheit. Auch wenn dabei die deutsche Betriebswirtschaftslehre zwar klar US-amerikanischen Entwicklungen folgt, so zeigen doch einige Aspekte immer noch kulturelle Eigenheiten von Deutschland bzw. Europa auf, z. B. im Teilgebiet der Experimental Finance. Der Beitrag wird abgerundet durch einige Betrachtungen, die die bis heute nicht immer einfache Lage der nicht-US-amerikanischen Forschung in einem traditionell US-dominierten Forschungsgebiet wie Finance beschreibt.
This paper examines an Epstein–Zin recursive utility with quasi-hyperbolic discounting in continuous time. I directly define the utility process supporting the Hamilton–Jacobi–Bellman (HJB) equation in the literature and consider Merton's optimal consumption–investment problem for application. I show that a solution to the HJB equation is the value function. The numerical and mathematical analyses show that unlike in the constant relative risk aversion utility, present bias in the Epstein–Zin utility causes economically significant overconsumption, maintaining a plausible attitude toward risks. Additionally, the sophisticated agent's preproperation occurs if and only if the elasticity of intertemporal substitution is larger than one.
Full-text available
Technological innovations drive the evolution of human societies. The success of innovations depends not only on their actual benefits but also on how potential adopters perceive them and how their beliefs are affected by their social and cultural environment. To deepen our understanding of socio-psychological processes affecting the new technology spread, we model the joint dynamics of three interlinked processes: individual learning and mastering the new technology, changes in individual attitudes towards it, and changes in individual adoption decisions. We assume that the new technology can potentially lead to a higher benefit but achieving it requires learning. We posit that individual decision-making process as well as their attitudes are affected by cognitive dissonance and conformity with peers and an external authority. Individuals vary in different psychological characteristics and in their attitudes. We investigate both transient dynamics and long-term equilibria observed in our model. We show that early adopters are usually individuals who are characterized by low cognitive dissonance and low conformity with peers but are sensitive to the effort of an external authority promoting the innovation. We examine the effectiveness of five different intervention strategies aiming to promote the diffusion of a new technology: training individuals, providing subsidies for early adopters, increasing the visibility of peer actions, simplifying the exchange of opinions between people, and increasing the effort of an external authority. We also discuss the effects of culture on the spread of innovations. Finally, we demonstrate that neglecting the cognitive forces and the dynamic nature of individual attitudes can lead to wrong conclusions about adoption of innovations. Our results can be useful in developing more efficient policies aiming to promote the spread of new technologies in different societies, cultures and countries.
Full-text available
Based on the literature on the relationship between culture, emotion, and loss aversion, we derive that culture can influence the degree of loss aversion. To test our hypotheses, we conduct a standardized survey in 53 countries worldwide that includes the questions from the Hofstede survey on cultural dimensions as well as lottery questions on loss aversion. The results show that individualism, power distance, and masculinity increase loss aversion as predicted, whereas the impact of uncertainty avoidance is less significant. Moreover, we also find a relation between the distribution of major religions in a country and loss aversion. In comparison, the connection of loss aversion to macroeconomic variables seems to be much smaller. Copyright
Full-text available
In order to assess the cross-cultural generality of monetary decision-making processes, American, Chinese, and Japanese graduate students were studied on two tasks: In the delay discounting task, participants made choices between immediate and delayed hypothetical monetary rewards; in the probability discounting task, participants made choices between certain and probabilistic rewards. Some notable cross-cultural similarities were observed. Two-parameter hyperbola-like functions described both delay and probability discounting for all three groups. Moreover, for all three groups the rate at which delayed rewards were discounted was higher for the smaller amount whereas the rate at which probabilistic rewards were discounted was lower for the smaller amount. Some group differences were also observed. As measured by the area under the empirical discounting curve, the Americans and Chinese discounted delayed rewards more steeply than the Japanese. In addition, the Americans discounted probabilistic rewards the most, whereas the Chinese discounted probabilistic rewards the least. Despite these differences, the similarities in the form of the discounting functions and in the effects of amount suggest that there are fundamental commonalities among the three groups with respect to the processes underlying their evaluation of delayed and probabilistic rewards.
Growth in this model is driven by technological change that arises from intentional investment decisions made by profit-maximizing agents. The distinguishing feature of the technology as an input is that it is neither a conventional good nor a public good; it is a nonrival, partially excludable good. Because of the nonconvexity introduced by a nonrival good, price-taking competition cannot be supported. Instead, the equilibrium is one with monopolistic competition. The main conclusions are that the stock of human capital determines the rate of growth, that too little human capital is devoted to research in equilibrium, that integration into world markets will increase growth rates, and that having a large population is not sufficient to generate growth.
This research explores whether there are systematic cross-national differences in choice-inferred risk preferences between Americans and Chinese. Study 1 found (a) that the Chinese were significantly more risk seeking than the Americans, yet (b) that both nationals predicted exactly the opposite - that the Americans would be more risk seeking. Study 2 compared Americans and Chinese risk preferences in investment, medical and academic decisions, and found that Chinese were more risk seeking than Americans only in the investment domain and not in the other domains. These results are explained in terms of a cushion hypothesis, which suggests people in a collectivist society, such as China, are more likely to receive financial help if they are in need (i.e. they could be cushioned if they fell), and consequently, they are less risk averse than those in an individualistic society such as the USA.