HLEG thematic workshop on Measuring Trust and Social Capital, Armin Falk

Nature and Predictive Power of Preferences: Global
Evidence
Armin Falk Anke Becker Thomas Dohmen
Benjamin Enke Uwe Sunde David Huﬀman
Institute on Behavior and Inequality (briq) and University of Bonn
Measuring Trust and Social Capital, Paris, June 10, 2016
Armin Falk Preferences: Global Evidence June 8, 2016 1 / 67

Motivation
Theories of individual decision-making take as fundamental a set of
preferences over risk, time, and social interactions
Empirical evidence generally scarce, scattered and restricted to a
limited set of countries, typically based on non-representative samples

What We Do
Measure and analyze global distribution of time and risk preferences,
positive and negative reciprocity, altruism, and trust
Tools
Preference module (Falk et al., 2015)
Novel, globally representative dataset (N=80,000) across 76 countries
(Falk et al., 2015)
Heterogeneity, determinants, predictive power

Outline
1 Global Preference Survey
2 Country-Level Descriptives
3 Preferences and Individual Characteristics
4 Preferences and Individual Behaviors
5 The Cultural Origins of Preference Variation
6 Outlook and Further Applications

Outline

Preference Module, Falk et al. (2015)
First comprehensive experimentally-validated preference survey
module
Leverages strengths of experimental and survey approaches
Two types of survey items were selected: Qualitative and quantitative
items (staircase)

Global Preference Survey
All items translated back and forth by professionals, into the major
languages of each target country
Monetary values for quantitative items adjusted along median
household income across countries
Pretests were conducted in 21 countries of various cultural heritage
Included in Gallup World Poll 2012
Representative samples in 76 countries, N=80,000

Global Preference Survey – Survey Items
For most of the six dimensions, underlying items comprise a combination
of quantitative and qualitative questions, set of 12 items
Preference Item Description Weight
Patience
Intertemporal choice sequence using staircase method 0.71
Self-assessment: Willingness to wait 0.29
Risk taking
Lottery choice sequence using staircase method 0.47
Self-assessment: Willingness to take risks in general 0.53
Positive Self-assessment: Willingness to return a favor 0.48
reciprocity Gift in exchange for help 0.52
Negative Self-assessment: Willingness to take revenge 0.37
reciprocity Self-assessment: Willingness to punish unfair behavior towards self 0.265
Self-assessment: Willingness to punish unfair behavior towards others 0.265
Altruism
Donation decision 0.54
Self-assessment: Willingness to give to good causes 0.46
Trust Self-assessment: People have only the best intentions 1
Table 1: Survey items of the GPS

In the Field

A
53-‐year
old
primary

school
teacher
in
rural

Cambodia
(Kandal

province)
during
the

interview

A
30-‐year
old
housewife

and
the
interviewer
in

rural
Bangladesh


Preference Module
Available in more than 100 languages, details in Falk et al. (2015) and on www.global-preferences.org (in progress)

Outline

World Map of Patience
Figure 2: World map of patience

World Map of Risk Taking
Figure 3: World map of risk taking

World Map of Positive Reciprocity
Figure 4: World map of positive reciprocity

World Map of Negative Reciprocity
Figure 5: World map of negative reciprocity

World Map of Altruism
Figure 6: World map of altruism

World Map of Trust
Figure 7: World map of trust

Country-Level Analysis
Most cross-country diﬀerences in preferences are statistically
signiﬁcant
t-tests of all possible (2,850) pairwise comparisons for each preference,
1-percent level
79% for risk, 82% for patience, 81% for altruism, 81% for positive
reciprocity, 79% for negative reciprocity, and 77% for trust
To which extent is this heterogeneity systematic?
Identify geographic and cultural patterns
Preferences correlated, giving rise to country-level preference bundles

Western and “Neo” Europe: vast majority of populations relatively
patient (10 most patient countries in sample all from this region);
average levels of risk taking; strong negatively reciprocal inclinations
Asia: Rather risk averse and impatient, except for the Confucian
countries (China, Japan, South Korea)
North Africa & Middle East: Relatively risk tolerant; low patience;
diverse social attitudes
Sub-Saharan countries: 10 most risk tolerant populations; all
countries below average in prosocial dimensions altruism, positive
reciprocity, and trust
Southern Americas: Rather impatient; low negative reciprocity;
intermediate risk taking

Patience Risk taking Positive reciprocity Negative reciprocity Altruism Trust
Patience 1
Risk taking 0.231∗∗ 1
Positive reciprocity 0.0202 -0.256∗∗ 1
Negative reciprocity 0.262∗∗ 0.193∗ -0.154 1
Altruism -0.00691 -0.0155 0.711∗∗∗ -0.132 1
Trust 0.186 -0.0613 0.363∗∗∗ 0.160 0.272∗∗ 1
Table 2: Preference outcome correlations
⇒ Patience and risk taking moderately correlated; high correlations among
“prosocial” traits altruism, positive reciprocity, and trust

Between- vs. Within-Country Variation
Preference
Between-country Within-country
variation (%) variation (%)
Patience 14.3 85.7
Risk taking 9.9 90.1
Positive reciprocity 11.4 88.6
Negative reciprocity 7.6 92.4
Altruism 12.3 87.7
Trust 8.3 91.7
Table 3: Between- vs. within-country variation

Outline

Preferences and Individual Characteristics
Investigate relationship of preferences to age, gender, and
(self-reported) cognitive ability; arguably exogenous
These characteristics are associated with important diﬀerences in
economic outcomes – suggests that preferences are part of the
explanation
Representative nature of data across countries allows investigation of
how general or culturally speciﬁc relationships are, e.g., between
gender and risk aversion

Preferences and Individual Characteristics
Dependent variable:
Patience Risk taking Pos. reciprocity Neg. reciprocity Altruism Trust
(1) (2) (3) (4) (5) (6)
Age 0.72∗∗∗ -0.083 1.02∗∗∗ -0.36∗ -0.0060 0.37∗
(0.17) (0.20) (0.17) (0.19) (0.14) (0.21)
Age squared -1.45∗∗∗ -1.20∗∗∗ -1.17∗∗∗ -0.45∗∗ 0.015 0.032
(0.20) (0.21) (0.18) (0.18) (0.15) (0.20)
1 if female -0.056∗∗∗ -0.17∗∗∗ 0.049∗∗∗ -0.13∗∗∗ 0.10∗∗∗ 0.066∗∗∗
(0.01) (0.01) (0.01) (0.01) (0.01) (0.01)
Subj. math skills 0.028∗∗∗ 0.046∗∗∗ 0.038∗∗∗ 0.040∗∗∗ 0.044∗∗∗ 0.056∗∗∗
(0.00) (0.00) (0.00) (0.00) (0.00) (0.00)
Constant -0.37∗∗∗ 0.21∗∗∗ -0.079∗∗ 0.37∗∗∗ -0.064∗∗ -0.078∗∗
(0.04) (0.04) (0.04) (0.05) (0.03) (0.04)
Country FE Yes Yes Yes Yes Yes Yes
Observations 78501 78445 78869 77521 78632 77814
R2 0.165 0.167 0.128 0.112 0.135 0.111
Table 4: Correlates of preferences at individual level. Age is divided by 100.

Preferences and Age
Figure 8: Age profiles by OECD membership. The figures depict the relationship
between preferences and age conditional on country fixed effects, gender, and
subjective math skills. Age is winsorized at 83 (99th percentile).

Universal and/or Country-Specificity: Gender effects across
countries (1/3)
Figure 9: Gender correlations separately by country. Green dots: not statistically
different from zero at the 10% level, while red / blue / pink dots denote countries
in which the effect is significant at the 1% / 5% / 10% level, respectively. Positive
coefficients imply that women have higher values in the respective preference.

Gender effects across countries (2/3)
-.4-.20.2.4
Coefficientonfemale
1 20 40 60 76
Countries
Pos. reciprocity
-.4-.20.2.4
1 20 40 60 76
Countries
Neg. reciprocity
Figure 10: Gender effects separately by country.

Gender eﬀects across countries (3/3)
-.4-.20.2.4
1 20 40 60 76
Countries
Altruism
-.4-.20.2.4
1 20 40 60 76
Countries
Trust
Figure 11: Gender eﬀects separately by country.

IQ effects across countries (1/3)
-.1-.050.05.1.15.2
Coefficientonmathskills
1 20 40 60 76
Countries
Patience
-.1-.050.05.1.15.2
1 20 40 60 76
Countries
Risk taking
Figure 12: Cognitive ability effects separately by country
Positive coefficients imply that higher IQ individuals have higher values in the respective preference.

-.1-.050.05.1.15.2
1 20 40 60 76
Countries
Pos. reciprocity
-.1-.050.05.1.15.2
1 20 40 60 76
Countries
Neg. reciprocity

-.1-.050.05.1.15.2
1 20 40 60 76
Countries
Altruism
-.1-.050.05.1.15.2
1 20 40 60 76
Countries
Trust

Income effects across countries (1/3)
-.04-.020.02.04.06
Coefficientonincomebracket
1 20 40 60 76
Countries
Patience
-.04-.020.02.04.06
1 20 40 60 76
Countries
Risk taking
Figure 15: Income effects separately by country
Positive coefficients imply that rich people have higher values in the respective preference.

-.04-.020.02.04.06
1 20 40 60 76
Countries
Pos. reciprocity
-.04-.020.02.04.06
1 20 40 60 76
Countries
Neg. reciprocity

-.04-.020.02.04.06
1 20 40 60 76
Countries
Altruism
-.04-.020.02.04.06
1 20 40 60 76
Countries
Trust

Outline

Preferences and Individual Behaviors
Investigate predictive power of preferences for economic and social
behaviors, around the world
Patience and accumulation decisions
Risk taking and risky choices
Social preferences and social interactions
Important to understand role of preferences in generating observed
variation in choice behavior...
... but also to provide an out-of-sample validation check on the
meaningfulness of the survey measures in culturally and economically
heterogeneous samples

Patience, Risk Taking, and Behaviors
Dependent variable:
Accumulation decisions Risky choices
Saved last year Education level Own business Plan to start business Smoking intensity
(1) (2) (3) (4) (5)
Patience 0.025∗∗∗ 0.033∗∗∗
(0.01) (0.00)
Risk taking 0.022∗∗∗ 0.017∗∗∗ 0.023∗
(0.00) (0.00) (0.01)
Constant -0.37∗∗∗ 0.24∗∗∗ -0.38∗∗∗ 0.038 0.39∗∗∗
(0.09) (0.07) (0.05) (0.03) (0.07)
Country FE Yes Yes Yes Yes Yes
Controls Yes Yes Yes Yes Yes
Observations 14459 69272 62985 51489 14490
R2 0.132 0.329 0.104 0.120 0.198
Table 5: Patience and accumulation decisions, risk preferences and risky choices

Social Preferences and Social Interactions
Dependent variable:
Donated Volunteered Helped Voiced opinion Have friends / relatives
money time stranger to oﬃcial I can count on
(1) (2) (3) (4) (5)
Altruism 0.061∗∗∗ 0.038∗∗∗ 0.052∗∗∗ 0.025∗∗∗ 0.016∗∗∗
(0.01) (0.00) (0.00) (0.00) (0.00)
Positive reciprocity -0.00037 0.0049 0.033∗∗∗ -0.0025 0.017∗∗∗
(0.00) (0.00) (0.00) (0.00) (0.00)
Negative reciprocity -0.0042 -0.0035 -0.0032 0.016∗∗∗ 0.0037
(0.00) (0.00) (0.01) (0.00) (0.00)
Constant -0.022 0.15∗∗∗ 0.37∗∗∗ -0.084∗∗ 0.48∗∗∗
(0.04) (0.05) (0.06) (0.03) (0.04)
Country FE Yes Yes Yes Yes Yes
Controls Yes Yes Yes Yes Yes
Observations 53439 53430 53226 53174 59209
R2 0.192 0.089 0.093 0.062 0.118
Table 6: Social preferences and social interactions

Outline

The Cultural Origins of Preference Variation
Preference variation may have deep cultural or historical roots
One approach: Use language structure as proxy for cultural variation
Chen (AER, 2013): Future-time reference (FTR) predictive of
future-oriented behaviors such as savings
Strong FTR: Explicit coding of future (I will go home); weak FTR:
coding of tense optional
Argument: Strong FTR makes future seem more distant, so people
less future-oriented

The Cultural Origins of Preference Variation
Our preference data allow two types of investigations
Replication and extension of Chens results with patience measure as a
more direct proxy for high future-orientation
Relationship between FTR and social preferences: if people care more
about the future, they should be more willing to invest in
relationships, build up reputations etc: data on positive reciprocity,
trust, and altruism ideally suited to investigate this hypothesis
Use interview language as proxy for household language and compute
country-level fraction of people speaking weak FTR languages

FTR and Preferences Across Countries
Dependent variable:
(1) (2) (3) (4) (5) (6)
Fraction of population speaking weak FTR 0.24∗∗∗ -0.0080 0.16∗∗ -0.088 0.099 0.18∗∗
(0.09) (0.07) (0.08) (0.09) (0.10) (0.08)
Log [GDP p/c PPP] 0.15∗∗∗ 0.032 -0.072∗ 0.058 -0.078∗ -0.00072
(0.04) (0.03) (0.04) (0.04) (0.04) (0.03)
Distance to equator 0.010∗ 0.0017 -0.0069 -0.0081 -0.0022 -0.0072
(0.01) (0.01) (0.01) (0.01) (0.01) (0.00)
Longitude -0.0019 0.0023 0.0022 0.00091 0.0025 -0.000039
(0.00) (0.00) (0.00) (0.00) (0.00) (0.00)
% at risk of malaria 0.25 -0.14 -0.27 -0.13 -0.71∗∗ -0.16
(0.21) (0.23) (0.29) (0.17) (0.27) (0.19)
Average precipitation -0.00024 -0.00092 0.00065 -0.0011 0.0031∗∗ -0.0011
(0.00) (0.00) (0.00) (0.00) (0.00) (0.00)
Constant -1.42∗∗∗ -0.041 1.39∗∗∗ -0.13 1.27∗∗ 0.56
(0.51) (0.36) (0.43) (0.50) (0.50) (0.38)
Continent FE Yes Yes Yes Yes Yes Yes
Observations 74 74 74 74 74 74
R2 0.636 0.442 0.253 0.271 0.334 0.408
Table 7: FTR and preferences at country level

Within-Country Evidence
In Switzerland, Estonia, and Nigeria, interview languages vary within
countries such that they generate variation in FTR (e.g., German vs.
English and French)
Within-country variation provides a powerful tool to identify cultural
origins of preferences because it allows us to keep unobserved country
characteristics ﬁxed
Use individual-level data with FTR assigned to interview language,
conditional on country FE and individual-level covariates

FTR and Preferences Within Countries
Dependent variable:
(1) (2) (3) (4) (5) (6)
1 if weak FTR 0.095∗∗ 0.067 0.18∗∗∗ -0.073∗ 0.24∗∗∗ 0.33∗∗∗
(0.05) (0.04) (0.04) (0.04) (0.04) (0.04)
Age 0.76∗∗∗ -0.098 1.07∗∗∗ -0.39∗∗∗ 0.032 0.42∗∗∗
(0.10) (0.10) (0.10) (0.10) (0.10) (0.10)
Age squared -1.51∗∗∗ -1.21∗∗∗ -1.22∗∗∗ -0.44∗∗∗ -0.017 -0.0044
(0.11) (0.11) (0.11) (0.11) (0.11) (0.11)
1 if female -0.061∗∗∗ -0.18∗∗∗ 0.045∗∗∗ -0.13∗∗∗ 0.100∗∗∗ 0.069∗∗∗
(0.01) (0.01) (0.01) (0.01) (0.01) (0.01)
Subj. math skills 0.028∗∗∗ 0.045∗∗∗ 0.039∗∗∗ 0.039∗∗∗ 0.044∗∗∗ 0.056∗∗∗
(0.00) (0.00) (0.00) (0.00) (0.00) (0.00)
Constant -0.49∗∗∗ 0.15∗∗∗ -0.13∗∗ 0.60∗∗∗ -0.17∗∗∗ -0.46∗∗∗
(0.06) (0.06) (0.06) (0.07) (0.06) (0.06)
Country FE Yes Yes Yes Yes Yes Yes
Observations 73460 73414 73811 72501 73580 72811
R2 0.166 0.172 0.127 0.117 0.137 0.113
Table 8: FTR and preferences at individual level

Outline

Outlook and Applications
First assessment of distribution and nature of preferences on a
globally representative basis using a novel dataset, which includes
behaviorally validated survey measures of preferences
Cross-cultural dimension of the data and the representative sampling
design allow entirely new perspectives and level of analysis:
Study relationship of preferences to individual characteristics in more
detail to understand biological or social mechanisms underlying, e.g.,
gender diﬀerences in preferences
Study sources of cross-country variation in preferences, or potential
co-evolution of preferences
Detailed investigation of link between aggregate outcomes and
preferences at the country level ⇒ Given previous lack of representative
preference data, new territory

Example Application: Patience and the Wealth of Nations
Example: Dynamic theories in both micro- and macroeconomics
highlight the crucial role of time preference for accumulation
processes and hence ultimately income
In neoclassical development framework, patience aﬀects accumulation
of physical capital (Ramsey-Cass-Koopmans), human capital (Becker,
Ben-Porath) as well as ideas and knowledge (Romer, Aghion &
Howitt)
Investigate consistency of this large and inﬂuential body of theoretical
work with empirical facts (Dohmen/Enke/Falk/Sunde, 2015)
Exploit variation in patience, accumulation, and income across
countries as well as across regions within countries

Patience and Contemporary Income
AFG
ARE
ARG
AUSAUT
BGD
BIH
BOL
BRA BWA
CAN
CHE
CHL
CHN
CMR
COL
CRI
CZE
DEU
DZA
EGY
ESP
EST
FINFRA GBR
GEO
GHA
GRC
GTM
HRV
HTI
HUN
IDN
IND
IRN
IRQ
ISR
ITA
JOR
JPN
KAZ
KENKHM
KOR
LKA
LTU
MAR
MDA
MEX
MWI
NGA
NIC
NLD
PAK
PER
PHL
POL
PRT
ROU RUS
RWA
SAU
SRB SUR
SWE
THA
TUR
TZAUGA
UKR
USA
VEN
VNM
ZAF
ZWE
4681012
Log[GDPp/cPPP]
-.5 0 .5 1
Patience
Patience and GDP
GEO
VNM
IRQ
ZWE
RUS
ZAF
CHL
PRT
HUN
BIH
LTU
ROU
GRC
EGY
THA
IRN
AFG
DZA
PER
HTI
GBR
NIC
BRA
PAK
IDN
UKR
HRVTZA
CAN
LKA
POL
MAR
KAZ
RWA
ITA
SAU
FRA
PHL
BWACRI
USA
BOL
JOR
CZE
VEN
GHA
CHN
IND
TUR
NGA
EST
GTM
MWI
DEU
KHM
COL
JPN
ARG
ESP
CMR
BGD
UGA
NLD
KOR
AUT
AUS
KEN
MEX
ARE
FIN
MDA
CHE
SWE
ISR
-2-1012
Log[GDPp/cPPP]
-.5 0 .5
Patience
(added variable plot)
Patience and GDP
Figure 18: Patience and per capita income. Added variable plot conditional on
geography, climate, continent FE, ethnic and gentic diversity, trust.
Similar results with short- and long-run growth rates since 1800, 1900,
1950

Patience and GDP
Holds within various sub-samples:
Within each continent separately
(Non-) OECD
(Not) colonized
Extends to other measures of development:
GDP per worker
Human development index (GDP, years of schooling, life expectancy)
Subjective life satisfaction

Patience and GDP
Robust to:
Controlling for inﬂation and interest rates
Including proxies for borrowing constraints
Restricting sample to top income quintiles

Patience and Accumulation
Dependent variable:
Human capital Physical capital TFP and Institutions
Schooling Educ. exp. Log [capital stock] Savings Log [TFP] R&D exp. Property rights
(1) (2) (3) (4) (5) (6) (7)
Patience 4.67∗∗∗ 1.45∗∗∗ 2.03∗∗∗ 7.70∗∗∗ 0.60∗∗∗ 2.09∗∗∗ 46.3∗∗∗
(0.53) (0.31) (0.28) (2.23) (0.10) (0.23) (4.39)
Constant 5.40∗∗∗ 4.28∗∗∗ 10.0∗∗∗ 10.2∗∗∗ -0.57∗∗∗ 0.96∗∗∗ 48.3∗∗∗
(0.24) (0.16) (0.13) (1.07) (0.05) (0.08) (1.95)
Observations 71 71 71 68 60 64 74
R2 0.429 0.138 0.327 0.102 0.265 0.574 0.515
Adjusted R2 0.421 0.125 0.317 0.088 0.253 0.567 0.508
OLS estimates, robust standard errors in parentheses. Correlations hold conditional on full set of covariates. ∗ p <
0.10, ∗∗ p < 0.05, ∗∗∗ p < 0.01.

Patience and Proximate Determinants
ARG
AUSAUT
BGD
BIH
BOL
BRA
BWA
CAN
CHE
CHL
CHN
CMR
COL CRI
CZE
DEU
EGY
ESP
EST
FIN
FRA
GBR
GEO
GHA
GRC
GTM
HRVHUN
IDN
IND
IRN
IRQ
ISR
ITA
JOR
JPN
KAZ
KEN
KHM
KOR
LKA
LTU
MAR
MDA
MEX
MWI
NGA
NLD
PAK
PER
PHL
POL
PRT
ROU
RUS
RWA
SAU
SRB
SUR
SWE
THA TUR
TZAUGA
UKR
USA
VEN
VNM
ZAF
ZWE
-9-8-7-6-5-4
Log[Capitalstockp/c]
-.5 0 .5 1
Patience
Patience and capital stock
Figure 19:

AFG
ARE
ARG
AUS
AUT
BGD
BOL
BRA BWA
CAN
CHE
CHL
CHN
CMR
COL
CRI
CZE
DEU
DZAEGY
ESP
EST
FIN
FRA
GBR
GHA
GRC
GTM
HRV
HTI
HUN
IDN
IND
IRN
IRQ
ISR
ITA
JOR
JPN
KAZ
KEN
KHM
KOR
LKA
LTU
MAR
MEX
MWI
NIC
NLD
PAK
PER PHL
POL
PRT
ROU RUS
RWA
SAU
SRB
SWE
THA TUR
TZA
UGA
UKR
USA
VEN VNM
ZAF
ZWE
0.005.0010.00
Averageyearsofschooling
-.5 0 .5 1
Patience
Patience and average years of schooling
Figure 20:

ARE
ARG
AUS
AUT
BGD
BIH
BOL
BRA
BWA
CAN
CHE
CHL
CHN
CMR
COL
CRI
CZE
DEU
DZA
EGY
ESP
EST
FIN
FRA
GBR
GEO
GHA
GRC
GTM
HRV
HUN
IDN
IND
IRN
ISR
ITA
JOR
JPN
KAZ KEN
KHM
KOR
LKA
LTU
MAR
MDA
MEX
MWINGA
NIC
NLD
PAK
PER
PHL
POL
PRT
ROU RUS
RWA
SAU
SRB
SWE
THA TUR
TZA
UGA
UKR
USA
VEN
VNM
ZAF
ZWE
0204060
Globalinnovationindex
-.5 0 .5 1
Patience
Patience and Innovation Index
Figure 21:

Relationship b/w patience and proximate determinants extends to
many other proxies for human and physical capital as well as factor
productivity
Holds for both stocks (years of schooling, capital stocks,...) and ﬂows
(savings, education expenditure as % of GDP,...)
Correlations hold conditional on full set of covariates...
... and often even conditional on GDP

Patience, Accumulation, and Income Across Regions
Dependent variable:
Log [Regional GDP p/c] Avg. years of education
(1) (2) (3) (4) (5) (6)
Patience 1.39∗∗∗ 0.21∗∗∗ 0.19∗∗ 3.34∗∗∗ 0.43∗∗ 0.40∗∗
(0.23) (0.07) (0.09) (0.55) (0.17) (0.16)
Constant 8.74∗∗∗ 9.18∗∗∗ 8.81∗∗∗ 7.17∗∗∗ 7.37∗∗∗ 6.97∗∗∗
(0.18) (0.02) (0.32) (0.36) (0.04) (0.55)
Country FE No Yes Yes No Yes Yes
Additional controls No No Yes No No Yes
Observations 704 704 687 693 693 676
R2 0.184 0.937 0.949 0.252 0.936 0.954
Adjusted R2 0.183 0.932 0.944 0.251 0.931 0.949
WLS estimates, observations weighted by number of observations in each re-
gion. Standard errors (clustered at country level) in parentheses. ∗ p < 0.10, ∗∗
p < 0.05, ∗∗∗ p < 0.01.

The Ancient Origins of the Global Variation in Risk
Preferences and Prosociality
What are the origins of substantial between-country variation?
One avenue: look at how historical events caused diﬀerences in
preferences (Becker/Enke/Falk, 2016)
This paper: human population pairs diﬀer in how long they have been
separated in the course of human history; e.g., Germans and Dutch
share relatively recent common ancestors, while the ancestors of
Germans and Chinese seperated tens of thousands of years ago

Hypothesis: The longer two populations have been seperated due to
long-run human migration patterns “Out of Africa” 50,000 years ago,
the larger their differences in preferences
For example, if experiences affect preferences, then longer separation
should induce larger preference differences because longer seperation
implies a larger number of idiosyncratic experiences

Empirical approach: form all possible country pairs, i.e., pair each
country with all other countries
Relate the absolute diﬀerence in a given preference between two
countries to the temporal distance between the respective
populations, i.e., proxies for the length of time since these populations
shared common ancestors (adjusted for post-Columbian migration)
Three classes of proxies: genetic distance (two types), linguistic
distance, theoretically predicted migratory distance measures (two
types) ⇒ ﬁve proxies

Main Result: the larger the temporal distance between two countries,
the larger their (absolute) difference in risk aversion, altruism, trust,
and positive reciprocity (prosociality)
Holds for five different proxies for temporal distance
Conditional on country FE, large vector of geographic, economic,
institutional controls
Takeaway: Differences in preferences across entire populations may
have VERY deep historical roots that go back thousands of years;
open question: channel (genetics, experiences and cultural learning,
random preference shocks)?

Thank you!

Patience and Longevity
Patience is strongly correlated with health proxies, both across and
within countries
Causality unclear: Patience might cause health, health might cause
patience, or income might cause both
But direction of causality is important to understand preference
formation as well as potential development traps:
High mortality risk ⇒ Low patience ⇒ Low investment ⇒ Low
income ⇒ High mortality
In standard intertemporal choice models with mortality risk, revealed
patience trivially declines in mortality risk

Patience is strongly correlated with health proxies, both across and
within countries
Causality unclear: Patience might cause health, health might cause
patience, or income might cause both
But direction of causality is important to understand preference
formation as well as potential development traps:
High mortality risk ⇒ Low patience ⇒ Low investment ⇒ Low
income ⇒ High mortality
In standard intertemporal choice models with mortality risk, revealed
patience trivially declines in mortality risk

Approach: Identify causal effect of longevity on patience by exploiting
differential change in life expectancy and patience across countries
and cohorts
Assign each individual their expected remaining years of life
conditional on country, age, and gender
Regression logic: if increase in life expectancy between a 20 and 40
years old Italian is larger than between a 20 and 40 years old German,
then the patience increase should also be larger; this approach nets
out unobservable country characteristics (GDP, institutions) as well as
aging patterns
Implementation: regress individual patience on life expectancy
conditional on country and age fixed effects, as well as lots of
individual-level covariates (gender, income, education, cognitive skills,
religion, etc)

Find signiﬁcant eﬀects
For both measures of time preference separately
Across several proxies for longevity
Holds up to controlling for institutional quality or national income,
averaged over lifetime

Further Details
Selected countries

Global Preference Survey: Selected Countries
East Asia and Paciﬁc: Australia, Cambodia, China, Indonesia, Japan,
Philippines, South Korea, Thailand, Vietnam
Europe and Central Asia: Austria, Croatia, Czech Republic, Estonia,
Finland, France, Georgia, Germany, Greece, Hungary, Italy, Kazakhstan,
Lithuania, Moldova, Netherlands, Portugal, Romania, Russia, Spain,
Sweden, Switzerland, Turkey, Ukraine, United Kingdom
Latin America and Caribbean: Argentina, Bolivia, Brazil, Chile, Colombia,
Costa Rica, Guatemala, Haiti, Mexico, Nicaragua, Peru, Venezuela
Middle East and North Africa: Algeria, Egypt, Iran, Iraq, Israel, Jordan,
Morocco, Saudi Arabia, United Arab Emirates
North America: United States, Canada
South Asia: Afghanistan, Bangladesh, India, Pakistan, Sri Lanka
Sub-Saharan Africa: Botswana, Cameroon, Ghana, Kenya, Malawi, Nigeria,
Rwanda,
South Africa, Tanzania, Uganda, Zimbabwe

HLEG thematic workshop on Measuring Trust and Social Capital, Armin Falk

Empfohlen

Empfohlen

Weitere ähnliche Inhalte

Ähnlich wie HLEG thematic workshop on Measuring Trust and Social Capital, Armin Falk

Ähnlich wie HLEG thematic workshop on Measuring Trust and Social Capital, Armin Falk (20)

Mehr von StatsCommunications

Mehr von StatsCommunications (20)

Kürzlich hochgeladen

Kürzlich hochgeladen (20)

HLEG thematic workshop on Measuring Trust and Social Capital, Armin Falk