ArticlePDF Available

National Intelligence Is More Important for Explaining Country Well-Being than Time Preference and Other Measured Non-Cognitive Traits


Abstract and Figures

Patient people fare better in life than impatient people. Based on this and on economic models, many economists have claimed that more patient countries should fare better than less patient countries. We utilize cross-national data in non-cognitive traits measured in the Global Preference Survey (GPS). This survey measured six non-cognitive traits — risk and time preferences, positive and negative reciprocity, altruism, and trust — across 76 countries in about 80,000 persons. As such, it provides the best current database of economics-focused non-cognitive traits. We combine this database with existing estimates of national intelligence (national IQs) and model country outcomes as a function of these predictors. For outcomes, we used the 51 national well-being indicators from the Social Progress Index (SPI) as well as the composite extracted from this, the general socioeconomic factor. We find that non-cognitive variables, time preference included, are only weakly predictive of national well-being outcomes when national IQs are also in the model. The median β across the indicators was 0.11 for time preference but 0.39 for national IQ. We replicated these results using six economic indicators, again with similar results: median βs of 0.15 and 0.52 for time preference and national IQ, respectively. Across all our results, we found that national IQ has 2-4 times the predictive validity of time preference. These results are fairly robust to inclusion of a spatial autocorrelation control, alternative measures of national IQ and time preference, or no controls. Our results suggest that the importance of national non-cognitive traits, including time preference, is overestimated or that these traits are mismeasured.
Content may be subject to copyright.
MANKIND QUARTERLY 2020 61:2 339-370
National Intelligence Is More Important for Explaining
Country Well-Being than Time Preference and Other
Measured Non-Cognitive Traits
Emil O. W. Kirkegaard*
Ulster Institute for Social Research, London, UK
Anatoly Karlin
Russian Academy of Sciences, Moscow
* Corresponding author:
Patient people fare better in life than impatient people. Based on
this and on economic models, many economists have claimed that more
patient countries should fare better than less patient countries. We
utilize cross-national data in non-cognitive traits measured in the Global
Preference Survey (GPS). This survey measured six non-cognitive traits
risk and time preferences, positive and negative reciprocity, altruism,
and trust across 76 countries in about 80,000 persons. As such, it
provides the best current database of economics-focused non-cognitive
traits. We combine this database with existing estimates of national
intelligence (national IQs) and model country outcomes as a function of
these predictors. For outcomes, we used the 51 national well-being
indicators from the Social Progress Index (SPI) as well as the composite
extracted from this, the general socioeconomic factor. We find that non-
cognitive variables, time preference included, are only weakly predictive
of national well-being outcomes when national IQs are also in the model.
The median β across the indicators was 0.11 for time preference but
0.39 for national IQ. We replicated these results using six economic
indicators, again with similar results: median βs of 0.15 and 0.52 for time
preference and national IQ, respectively. Across all our results, we
found that national IQ has 2-4 times the predictive validity of time
preference. These results are fairly robust to inclusion of a spatial
autocorrelation control, alternative measures of national IQ and time
preference, or no controls. Our results suggest that the importance of
national non-cognitive traits, including time preference, is overestimated
or that these traits are mismeasured.
Key Words: Intelligence, Time preference, Patience, Trust, Social
Progress Index, Human Development Index, Non-cognitive traits
Patience, also called low time preference, future time orientation, delayed
gratification and delay discounting, has long been associated with higher
intelligence and socio-economic outcomes. In a famous 1972 experiment,
Stanford psychologist Walter Mischel tested 82 young children on their ability to
delay the gratification of a small reward in the here and now to get a larger reward
15 minutes later (Mischel, Ebbesen & Raskoff Zeiss, 1972). Later studies would
find correlations between children who passed this “marshmallow test” and their
performance on SAT scores, educational attainment, body mass index (BMI), and
other indicators of life success. In a 1990 study, it was found that the “impulse
controlled” third of students who waited scored 610 in verbal and 652 in math,
versus 524 in verbal and 528 in math amongst the “impulsive” third who gobbled
up the marshmallows (Mischel, Shoda & Rodriguez, 1989).
A recent large replication, however, found much weaker results: the
“bivariate correlation [between marshmallow performance and outcomes] was
only half the size of those reported in the original studies and was reduced by two
thirds in the presence of controls for family background, early cognitive ability,
and the home environment”, and “we observed that delay of gratification was
strongly correlated with concurrent measures of cognitive ability, and controlling
for a composite measure of self-control explained only about 25% of our reported
effects on achievement.” (Watts, Duncan & Quan, 2018). As such, this famous
experiment joins many other prominent studies that have recently experienced
replication or effect size decline issues (Bavel & Cunningham, 2017; Suter &
Suter, 2019). However, even if the marshmallow test turns out to be not as
predictive as once thought1, it is still not unreasonable to expect that patience is
adaptive so far as positive life outcomes are concerned.
1 Indeed, it would be amazing if it was very predictive since psychological traits measured
at age 4 are usually not that strongly correlated even with adult trait levels owing to
developmental speed differences. Intelligence measured with the Stanford-Binet at age
4 correlates about .49 with that measured at age 17, so we should not be surprised that
a relatively weak and brief measurement such as the marshmallow test does not have
impressive predictive validity for late teen or adult life outcomes (Jensen, 1980, p. 279).
The pattern of delayed gratification correlating with higher intelligence
appears to extend to the animal kingdom. For instance, it is the smartest birds
such as corvids and Gray African parrots that have been found to display the most
impressive levels of self-control, waiting up to 15 minutes for a tastier treat
(Hillemann et al., 2014; Koepke, Gray & Pepperberg, 2015). One study found that
self-control and intelligence were positively related in Chimpanzees (Beran &
Hopkins, 2018).
Differences in patience between countries may have deep historical roots
(Galor & Özak, 2016). Economic historian David Landes dates it to as far back as
the invention of the mechanical clock in 13th century Europe, which he associated
with a macro-psychological transition from the “task-oriented time consciousness
of the peasant (one job after another, as time and light permit)” to the modern
“effort to maximize product per unit of time (time is money)” (Landes, 1998). One
important effect of the European clock-making tradition was to reinforce European
dominance in seafaring, since accurate clocks allowed for good estimates of
longitude (Sobel, 2010). Notably, while both China and the Islamic world
displayed an interest in these West European contraptions, the mechanical clock
would not assume a central place in their public spaces until the modern age;
doing so, as Landes argues, would have challenged the sovereignty of the
Emperor and the muezzins over time. Curiously, even today it is Protestant
Europe the medieval birthplace of the mechanical clock that continues to
have the most accurate clocks in public spaces, with Switzerland in absolute first
place (Levine & Norenzayan, 1999).
Unsurprisingly, the link between high levels of patience and socio-economic
success has also been found at the national level. A recent large study of 80,000
people across 76 countries found positive relationships of time preference with
several other personality variables and indicators such as individual savings
decisions, labor market choices, and prosocial behaviors (Falk et al., 2018). In an
earlier study of 53 countries, in which participants were asked whether they would
prefer to take $3,400 now or $3,800 later, lower time preferences were associated
with higher levels of technological innovation and environmental protection
(Wang, Rieger & Hens, 2016).
However, as we saw, there is good reason to believe that low time preference
is a medium to strong correlate of intelligence, especially at the group level
(Jones, 2012; Shamosh & Gray, 2008). This opens the possibility that the
observed time preference associations with economic performance for countries
are due to confounding with intelligence, and do not reflect the causal effects of
time preferences themselves, or at least, not mainly so. Meanwhile, numerous
studies have demonstrated that national average intelligence (national IQ) is a
strong correlate of a nation’s GDP per capita, innovation rates, standards of living,
and most other measures of country well-being (Lynn & Becker, 2019;
Rindermann, Kodila-Tedika & Christainsen, 2015). In particular, one can factor
analyze a collection of such country indicators to extract a single overall indicator
of a country’s well-being, akin to the Human Development Index (HDI). National
IQ is very strongly correlated with this indicator, r = .87, no matter the factor
analysis method, the particular indicators included, or the data source
(Kirkegaard, 2014).
Importantly, many economists have examined the relationship for causality
and generally found that it is supported. Evidence for this comes from tests for
backwards causation, where it is found that national IQ predicts later economic
outcomes, even controlling for earlier economic outcomes and plausible
confounders such as geography, legal traditions and so on (Christainsen, 2013,
2020; Hanushek & Woessmann, 2012; Jones & Potrafke, 2014; Jones &
Schneider, 2006; for a review of all studies, see Kirkegaard, 2020). Considering
these findings, we are obliged to ask: do time preferences still play an important
role in predicting country well-being outcomes when we include national IQ
scores in the regression models? To examine this question was the purpose of
the present study.
National intelligence
National IQ data was extracted from the book Intelligence: A Unifying
Construct for the Social Sciences (Lynn & Vanhanen, 2012). Although Lynn’s
datasets have been subjected to criticism, scholarly opinion has generally
changed from negative to positive and many independent research teams now
use them (Dutton, van der Linden & Madison, 2019; Grigoriev & Lapteva, 2018;
Hafer, 2016; León, 2018; Lv & Xu, 2016; Minkov, Welzel & Bond, 2016; Nikolaev
& Salahodjaev, 2016; Rindermann & Becker, 2018; Thies, 2019). More
importantly, a recent recalculation of the national IQs from the original sources by
David Becker has validated the original calculations. Not only were there relatively
small discrepancies between the results, but measures of bias on Lynn’s part
produced near-zero results.2 The new dataset is continuously updated at (current version v1.3.3), and has also been made available in book
form which provides an updated review of the various published findings related
2 These were done by computing the discrepancy scores between Lynn’s results and the
redone values, and seeing whether these correlated with variables. They did not (David
Becker, personal communication, 2018).
to national IQ (Lynn & Becker, 2019). However, even though the Becker dataset
is more methodologically rigorous, it covers fewer studies thus far (about 70% of
the Lynn 2012 database), and thus has less reliable and fewer estimates
(Kirkegaard, 2019b). Thus, in some ways, the 2012 Lynn dataset is still superior
for empirical analysis. In light of the continued debate about these data, we also
included the values computed by Heiner Rindermann (Rindermann, 2018), and
those of the World Bank (
scores, Angrist et al., 2019) as further alternatives. Rindermann’s values are
based mainly on the scholastic tests (Rindermann assigned them 3x weights
compared to the IQ data), whereas Lynn and Becker’s are based on IQ tests and
scholastic tests in more equal weights. The World Bank data are solely based on
the scholastic tests.
Non-cognitive traits
As mentioned earlier, there were two recent global studies that collected time
preference data for countries. The main data source is the Global Preference
Survey (GPS) (Falk et al., 2018). The GPS measured non-cognitive traits across
76 countries in about 80,000 persons (
preferences/home). This was a massive and expensive undertaking (1.3 million
Euro from the EU, Time preference
data were also available from another source, Wang, Rieger & Hens (2016), who
relied on ad hoc, sometimes small samples from 53 countries to estimate time
preferences across the world. The GPS survey is to be preferred for multiple
reasons. First, it included better sampling of countries, covering 76 of them.
Second, the sampling within countries was much stronger in Falk et al. who used
large and representative samples while Wang et al. relied on ad hoc student
samples. The median sample size in Wang et al. was 100 (median absolute
deviation = 55), while in Falk et al. the median was about 1000. Third, Falk et al.
also took five other non-cognitive (cultural or personality) measures, which can
be included as covariates: risk taking, positive reciprocity, negative reciprocity,
altruism, and trust3. The authors of the Falk et al. dataset went to considerable
lengths to ensure that their measures worked as expected. The details of this
3 These are derived from the various economic games that economists have people play.
For instance, “For negative reciprocity, we conducted two different experiments. A
subject’s minimum acceptable offer in an ultimatum game serves as one assessment
of negative reciprocity. We obtain a second assessment from a subject’s investment
into punishment after unilateral defection of their opponent in a prisoner's dilemma.”
(Falk et al., 2016)
work can be found in Falk et al. (2016). Figure 1 shows a world map of the time
preference measure in the GPS.
A reviewer pointed out that a recent working paper had meta-analyzed 10
datasets of time preferences data, including both Wang et al. and Falk et al., and
produced a more comprehensive dataset (n = 114 countries), though with many
countries having poor quality data (Rieger, Wang & Hens, 2020). We included
their data as an additional measure of patience, but used Falk et al’s as our main
dataset due to having the highest coverage and using the same method for all 76
countries, thus ruling out method variation.
Figure 1. World map of standardized time preference based on data from Falk
et al. (2018). Green = low time preference / high patience; Red = high time
preference / low patience.
Well-being measures
For our measures of country well-being, we primarily used the data from the
2019 Social Progress Index dataset (SPI, This
is an index based on 51 diverse indicators and is marketed as a “beyond GDP”
alternative to the popular GDP per capita measure and the United Nations’
Human Development Index (HDI,
development-index-hdi), which relies on GNI4 per capita as one of its three
4 GNI = Gross National Income. GDP = Gross Domestic Product. “GDP is the total market
value of all finished goods and services produced within a country in a set time period.
components. The indicators in the SPI comprise measures of health (e.g.
stunting), sanitation (e.g. access to drinking water), modern infrastructure (e.g.
electricity), crime (e.g. homicide rate), education (e.g. primary school enrollment),
information society (e.g. internet access), political system (e.g. freedom of
speech), governance (e.g. corruption), and various others. The SPI however does
not cover as many countries as the HDI, so we included this measure as a
secondary outcome variable (2018 edition). For the SPI dataset, we calculated
each country’s score on the general factor extracted from the component
variables as done in a prior study (Kirkegaard, 2014). The reason for doing this is
that this composite better captures the common variance in the components
compared to the theory-based measure the authors advocate.5 The general
factor, called S (for socioeconomic), ‘explained’ 49% of the variance across the
51 indicators. It is important to note that this general factor should not necessarily
be understood as a reflective factor, i.e. a factor that causally affects the 51
indicators. Rather, we understand the S factor to be a weighted index of many
correlated and causally interrelated social variables. Whether one understands
this as a formative factor or a convenient summary of a causal network is mostly
irrelevant for the purposes of this study (on the relationship between factor
analysis and network models, see Christensen & Golino, 2020). Figure 2 shows
the approximate causal model assumed in this study.
In this simplified model, there exists a large variety of socioeconomic
variables in a complex reciprocal causal network with each other [in the box].
Outside influences on this network come mainly from intelligence and non-
cognitive traits. There is an additional impact of contextual factors such as
geography, historical dependencies (e.g. legacy of communism, civil war), natural
resources (e.g. oil) and so on. The “deep roots” causes of psychological variation
is chiefly demographic variation, which ultimately implies genetic variation
(Fedderke et al., 2014; Murphy & Nowrasteh, 2016; Spolaore & Wacziarg, 2013).
It is not our goal to argue for this complete model in this paper, merely to explain
our larger theoretical approach. The results of this study are not dependent on
this deep demographic roots interpretation of between country variation in
intelligence and non-cognitive traits.
GNI is the total income received by the country from its residents and businesses
regardless of whether they are located in the country or abroad.”
5 Which seems to be based on no serious theory that we can find, but seems to reflect
mainly how the authors categorize the various indicators based on intuition.
The SPI database does not contain any economic measures. This stands in
stark contrast to prior research on country outcomes which is strongly focused on
GDP per capita and GDP growth. To offset this deliberate sampling bias in the
SPI, we added a few economic outcomes as well. Specifically, we downloaded
GNI and GDP per capita data from the World Bank for the years 1990-2019
(, We supplemented
this with median income data from Gallup. The Gallup data are based on self-
reported median income data, for individuals and households in the period 2006-
2016, aggregated by them (
median-household-income-000.aspx). We operationalized economic growth as
log10(value 2019/value1990). The other economic variables were likewise log10
transformed, as is custom to reduce the skew.
Figure 2. Approximate causal model of country well-being and its causes.
Other variables
We included regional dummies (labeled macroregions) based on the United
Nations’ classification, which we modified to reduce the number of dummies to a
more manageable level (n = 14). This level of dummies is intermediate between
continents (n = 5) and UN regions (n = 23) and has been used in prior studies
(Kirkegaard, 2019a). The supplementary materials show maps of the
classifications used (Figure S1-S3). The purpose of including regional dummies
as predictors is that they capture validity of any omitted variables that relate to
regional breakdown of the world. This can be thought of as vague cultural
preferences or norms, languages, economic conditions and so on. This also
partially captures spatially autocorrelated omitted variables.
We used global geodata from
world_borders.php. This was used for plotting purposes and for calculation of
spatial lag variables. We calculated spatial lag variables by averaging the values
of the three nearest neighboring countries (in the full dataset) as measured by the
great circle distance between their capitals. While one could use another number
of neighboring countries than three, this value has been found to work well in prior
research (Fuerst & Kirkegaard, 2016; Kirkegaard, 2016).
We used more restrictive p value thresholds with a default ‘statistical
significance’ of .01. This value was chosen prior to the modeling in order to
decrease the false positive rate in line with a recent proposal (Benjamin et al.,
2018). We did not go as far as the authors suggest (to .005) because our use of
a country level dataset would mean that such statistical certainty is very rarely
met. Thus, our chosen level of certainty is a compromise between data limitations
and limiting false positives, which is however, 5 to 10 times stricter than the
commonly used standards .05 and .10.
Social well-being overall
Before turning to the main results, we looked at the issue of range restriction
in the sample with data coverage. While the datasets used in this study cover a
large fraction of the world’s population, the included countries are not a random
subset of all countries, but are rather somewhat selected for being countries with
better government statistics and ease of doing large surveys. This selection bias
induces range restriction in the models since the data for many poor countries
are not observed. To examine the degree to which this was the case, we
computed the standard deviation in our full sample for national IQs (Lynn 2012
dataset) and the SPI-based S factor, as well as in the subsets of the countries
with data for the non-cognitive traits. Table 1 shows the results.
Table 1. Means ± standard deviations and sample sizes N by subsets of the
dataset used in the study.
Full dataset
Falk et al.
Wang et al.
Rieger et al.
84.3 ± 10.9
88.1 ± 10.3
94.4 ± 7.8
88.3 ± 10.3
0.00 ± 1.00
0.31 ± 0.85
0.82 ± 0.73
0.38 ± 0.85
196, 176
The results show that the subsets used in the study are somewhat selected.
The Falk et al. dataset only has about 10% range restriction in the outcome
variable, but less in the national IQ variable. The Wang et al dataset has more
severe selection bias, with 27-29% reduced SD in both variables. With this range
restriction issue in mind, Table 2 shows the correlation matrix of the main
variables used in this study.
Table 2. Correlation matrix of main variables. Values below the diagonal are
weighted by the square root of population size. TP = time preference, lv = Lynn
and Vanhanen, b = Becker, r = Rindermann, wb = World Bank. * = p < .01, ** = p
< .005, *** = p < .001.
0.88*** 0.86*** 0.90*** 0.93*** 0.94*** 0.63***
TP Wang 0.61*** 0.60*** 0.53*** 0.62*** 0.62*** 0.66***
TP Falk
0.61*** 0.62*** 0.63*** 0.63*** 0.56*** 0.59*** 0.57***
TP Rieger
Pos. recip.
0.37*** 0.30* 0.36** 0.28 0.19 0.21 0.21
Neg. recip.
Table 2. Correlation matrix of main variables. Values below the diagonal are
weighted by the square root of population size. TP = time preference, lv = Lynn
and Vanhanen, b = Becker, r = Rindermann, wb = World Bank. * = p < .01, ** = p
< .005, *** = p < .001.
Altruism Trust
TP Wang 0.63*** 0.82*** -0.21 0.05 -0.03 -0.17 0.22
TP Falk
0.84*** 0.13 0.07 0.22 -0.05 0.17
0.19 0.01 -0.26 0.19 -0.02 -0.06
0.13 0.08 -0.21 -0.15 0.71*** 0.36**
0.20 0.22 0.14 -0.17 -0.13 0.16
Trust 0.25 0.12 0.03 0.56*** 0.13 0.47***
As found in much previous research, the weighted and unweighted results
are quite similar. Our preferred specification is to use the weighted results (Fuerst
& Kirkegaard, 2016). We note that the country well-being measures have stronger
correlations to national IQ estimates (IQlv and IQr, mean r = .77) than to the time
preference estimates (mean r = .60) and the other variables in Falk et al.’s dataset
(r’s from -.14 to .19). Thus, prima facie, these other variables seem unlikely
candidates to explain variation in national well-being. Furthermore, the World
Bank’s IQ estimates, which they label “Human Capital Index”, are more strongly
correlated with the SPI S factor than the other IQ measures. The reason for this
is that the World Bank utilizes test scores on scholastic tests, but also
incorporates data about health and schooling length into their measure (Angrist
et al., 2019; World Bank, 2018) since their goal is to “capture the amount of
human capital a child born today could expect to attain by age 18.” Because of
this, their measure is a mix of cognitive and non-cognitive data. Thus, we do not
use it further here.
Figures 3-5 show the scatterplots between national IQ, Falk et al.’s patience
measure, and the S factor index of well-being.
Figure 3. Scatterplot of national IQ (Lynn & Vanhanen, 2012) and time
preference (Falk et al., 2018), weighted by the square root of population size.
Figure 4. Scatterplot of national IQ (Lynn & Vanhanen 2012) and the S factor
(broad well-being index). Weighted by the square root of population size.
Figure 5. Scatterplot of time preference (Falk et al., 2018) and the S factor (broad
well-being index). Weighted by the square root of population size.
Figure 3 shows that while there is a strong positive correlation between
national IQ and time preference, the North and West European countries
constitute a group of positive outliers. This is the same area where the mechanical
clock was first introduced and where bank clocks are the most accurate, as
mentioned in the introduction. Next we moved to regression analysis to examine
the ability of time preference to predict country well-being in the presence of
national IQ and other controls. Table 3 shows the main regression models.
We see that across the 8 models, national IQ never has a standardized β
below 0.43, 7 of the 8 βs have p < .01, the last has p = .026. Thus, by conventional
standards, national IQs are fairly robust predictors, in line with prior research
(Jones & Potrafke, 2014; Jones & Schneider, 2006). Time preference, on the
other hand, is a decent predictor when alone (model 3, β = 0.48). However, once
it is in models together with national IQ and various controls, it is weak, β -0.06 to
0.16, and does not even reach p < .05 in the five models that feature IQ as well
(models 4-8). Thus, time preference was not generally a useful predictor beyond
its association with national IQ. The results are even worse for the other five non-
cognitive measures. The first four of them never reach p < .05, and have near-
zero βs, just as they had near-zero correlations to begin with. Trust, however, is
consistently negative at about -0.20. Of the three βs, all p’s are < .05, and two are
below .01. Thus, the results suggest the improbable conclusion that holding
national IQ, time preference and the other non-cognitive traits constant, as well
as the geography controls, trust has a negative influence on overall social well-
being. With regards to the geographical dummies, we may observe that they are
all negative in sign, and some of them large and with small p values. The index
region is Northern & Western Europe, so the results here indicate that even in
models that account for intelligence and 6 non-cognitive traits, there are still
residual negative geographical effects associated with being outside NW Europe.
This would suggest that this region enjoys some advantages that are not properly
measured by the included predictors. The effect size of these is quite large. Model
7, which introduces the geographical variables, explains about 20% more
variance than model 6.
Table 3. Main regression model results. Outcome = Social Progress Index
general factor (S factor). Standardized β is shown with standard error in
parentheses. lv = Lynn & Vanhanen, b = Becker, r = Rindermann, wb = World
Bank. Time pref. = Falk et al. time preference. Weighted by the square root of
population size. Model 2 is identical to model 1 but is fit using only the countries
that have data in Falk et al.’s dataset. Values in parentheses are standard errors.
* = p < .01, ** = p < .005, *** = p < .001.
Time pref.
Latin America
Central Asia
Eastern Asia
Eastern Europe
Southern Asia
Southern Europe
Pos. recipr.
Neg. recipr.
Spatial lag
R2 adj.
To test the robustness of our results, we also ran variations on these models
which are given in full in the supplementary materials. First, we ran models using
the time preference measure from Wang et al., as well as a combined version of
the two time preference sources (combined by averaging scaled Z scores, Table
S1 and S2). Second, we ran models using HDI as an alternative outcome (Table
S3). Third, we ran the regressions using the alternative measures of national
intelligence produced by Heiner Rindermann, and by David Becker (Tables S4-
S5). Fourth, we ran the main models without population size weights (Table S6).6
All of these alternative approaches produced fairly similar results to the ones
shown in Table 3 above. Table 4 shows β estimates from the models concerning
the two primary predictors of interest. These are all based on the final, most
inclusive model (analogous to model 8 in Table 3).
Results from the robustness tests showed that national IQ was generally but
not entirely a robust predictor. The β of IQ ranged from 0.24 to 0.64, with a median
of 0.45 compared to the 0.46 found in the main model specification (i.e. from
Model 8 in Table 3, above). To some degree, weaker results could be interpreted
as resulting from poorer data or less representative datasets. The result based
on Wang et al.’s data had a reduced sample size of 51 compared to our main
model with 76, and as we showed in Table 1, the Wang dataset results in
moderately strong range restriction. The other outlier in terms of the β for IQ was
6 We also reran the primary models using robust regression (Table S7). This also
produced similar results. We did not include these in the summary because the model
functionality in R did not allow for calculation of R².
Becker’s dataset. Inspection of the models in Table S5 revealed that the decrease
in IQ’s β was due to the inclusion of the regional dummies in the models.
Seemingly, Becker’s IQ calculations have poorer within continent discrimination
than Lynn & Vanhanen’s and Rindermann’s. In contrast to our IQ results, the β
for time preference was only slightly positive across all models, with a median β
of 0.12, ranging from 0.08 to 0.21. Thus, this variable does not appear to have
much validity to predict country well-being. As noted in our discussion of the
primary model results, trust appears to have a generally negative β. This was also
true across the supplementary models, where it had a median β of -0.16, ranging
from -0.20 to -0.11. Risk-taking had a median β of 0.10 with a range from 0.01 to
0.14, so might have some validity. The remaining variables had near-zero values.
Table 4. Selected results from alternative model specifications from the models
in Table 3 and Tables S1-S6. Full regression output including standard errors and
p values can be found in the supplementary materials.
Model approach
+ Falk
+ Falk
+ Falk -
Spatial lag
Time preference
Positive recipr.
Negative recipr.
With regards to spatial autocorrelation, the use of the UN macroregional
dummies seemed sufficient. Substitution of the spatial lag variable for the regional
dummies did not improve model fit in any case, in fact, reduced it somewhat.
Figure 6 shows the model adjusted R2s.
We see that in every case, the models with just national IQs outperform those
with just time preference by a large margin (models 1 vs. 2). The models that
include both predictors do a little better than those with just national IQs. We
conducted a likelihood ratio test on the inclusion of the time preference predictor
beyond the national IQ (i.e., models 1 vs. 3). The p values from these tests were:
.061, .015, .006, .037, .141, .035, .007. Thus in 2 of the 7 cases were the p values
below this study’s threshold of .01. Thus, there is some evidence that time
preference has incremental validity, but it is not impressive.
Figure 6. Model adjusted R2’s across specifications.
Indicators of the Social Progress Index
Moving beyond the SPI S factor, we next studied each of the 51 indicators in
the SPI database. As mentioned earlier, this is a set of very diverse country-level
indicators. It has been found that such indicators generally are positively related
such that good things go together. Some of the indicators are coded in the
negative direction. These measure something considered undesirable, such as
the murder rate. To accommodate these, we coded the outcome by reversing the
indicators with negative loadings on the S factor in the factor analysis. We fit the
eight models from Table 3 for each outcome variable and saved all the results
(408 models total). To summarize the results across these, we extracted the full
model with regional dummies and compared the β’s for national IQ and time
preference for each outcome, shown in Figure 7.
We see that across models, national IQ is clearly the more important
predictor, just as we saw for the main results for the S factor. Quantitatively, the
median/SD β’s are 0.39/0.21 for national IQ and 0.11/0.12 for time preference.
Thus the mean national IQ β is 3.6 times as great as that for time preference. If
we do this ratio comparison within models and compute the median of the ratios,
we get a value of 3.5. If we instead compare the β’s from the direct comparison
models (model 4, no covariates), we get materially the same results: median β’s
of 0.40 and 0.14, thus with a ratio of 2.9. If we use the within model ratios, the
median is 3.0. Thus, we find that the choice of outcome variable is not important
for our general findings, and neither is the exact method used to compare the
relative importance of the two variables.
Figure 7. Comparison of standardized betas (β) for national IQ and time
preference across 51 indicators found in the Social Progress Index database.
Economic outcomes
As mentioned, the designers of the SPI eschew economic variables. To
counter this selection bias in outcome variables, we collected varied economic
variables ourselves. We employed the same modeling approach for the economic
variables as for the SPI indicators. Figure 8 shows the pairwise scatterplots and
distributions of the economic variables.
Next we fit the same set of regression models to each of the 6 economic
outcomes as done in the SPI section above. Figure 9 shows the βs from the full
model with regional dummies. The plot shows that in each case, national
intelligence predicts the outcome better than does time preference, though all β’s
are positive. The median β’s are 0.52 and 0.15, for national IQ and time
preference, respectively, yielding a ratio of 3.5. Likewise, if we compute the
median ratio within models, it is 1.8 (though the mean is 7.7!). In the models
without controls, the median β’s are 0.54 and 0.19, for national IQ and time
preference, yielding a ratio of 2.8, and the median ratio is 2.9. Thus, our results
for the economic outcomes mirror those for the non-economic outcomes.
Figure 8. Pairwise associations of economic variables along with national
intelligence (Lynn & Vanhanen, 2012) and the first factor of the Social Progress
Index variables (S factor). hh = household. Unweighted correlations.7
7 The function for this plot did not support the use of regression weights. The weighted
correlation matrix can be found in the supplementary materials.
Figure 9. Comparison of standardized betas (β) for national IQ and time
preference (patience), across 6 economic outcome measures.
Thousands of published papers emphasize the importance of patience in life,
whether this is conceptualized as having a weak preference for time (low time
preference, in economists’ terms), or being higher in conscientiousness (in
psychologists’ terms). There is ample evidence that at the individual level,
measures of patience predict better life outcomes, ranging from job success to
health. Some of this validity is retained when measures of intelligence are
included in models (O’Boyle et al., 2011; Schmidt, Oh & Shaffer, 2016; Zissman
& Ganzach, 2020). Thus, it is not surprising to find that researchers who study
global variation in country well-being, whether focused on economic outcomes
such as GDP per capita or something else, have attempted to explain these in
terms of patience differences between populations (Dohmen et al., 2015, 2018;
Falk et al., 2018; Galor & Özak, 2016; Wang, Rieger & Hens, 2016).
As we mentioned earlier, however, the evidence indicates that at the
individual level, patience is positively associated with intelligence, and thus
models that leave out intelligence are likely to spuriously assign some of its
validity to patience measures. Indeed, this was found to be the case for the much
discussed marshmallow test (Watts, Duncan & Quan, 2018). Two recent studies
by economists using the same dataset of time patience as the present study found
that this is correlated with ‘human capital’ and intelligence as measured from
various test scores (Hanushek et al., 2020; Potrafke, 2019). From another
research perspective, researchers who have studied the importance of
intelligence (or ‘cognitive ability’) have argued there are clearly other important
factors between populations that explain variation in country outcomes, aside
from intelligence and geographic factors such as natural resources (Christainsen,
2020). They too have speculated that such might be patience, creativity, work
ethic and so on (Lynn & Vanhanen, 2012; Rindermann, 2018).
Thus, we are left with the important question: How well do time preferences
predict socioeconomic well-being when we take intelligence into account? We
studied this question by comparing models that include time preference,
intelligence, 5 other non-cognitive traits, as well as geographical controls. Our
study used the 51 indicators of the Social Progress Index (SPI), the composite
overall well-being score extracted from this database, as well as 6 economic
variables. Our study of 51+6 indicators of national well-being in a single analysis
is probably the most comprehensive national IQ validity study to date. Most prior
studies have relied only upon a few indicators, especially the Human
Development Index and various economic outcomes such as GDP per capita
(nominal, PPP, and variants: GNI per capita, and so on).
We find that overall, national IQ is a better predictor of outcomes than (low)
time preference as well as the five other non-cognitive traits measured by the
Global Preference Survey (risk-taking, positive reciprocity, negative reciprocity,
altruism, and trust). We find this result across hundreds of regression models that
include variation in the inclusion of controls, different measures of time
preference, and different outcomes. Thus, our results appear quite robust. Our
results do show some evidence of time preference’s positive validity, but it is fairly
marginal, sometimes having a small p value in one model but not in the next.
Based on the studies of individuals, we interpret this as showing that time
preference probably has some positive effects between countries, but that we do
not have sufficient statistical power here to consistently spot them.8 The most
surprising finding of our regression models is that trust generally had negative
coefficients, although generalized trust is generally regarded as a public good and
8 For an analogous case, consider the history of the so-called specificity hypothesis in
industrial and organizational psychology (Salgado, 1995; Schmidt & Hunter, 1984).
is strongly correlated with other positive indicators in subnational analyses (Carl,
2014). We have no good explanation for this result, though it was not entirely
robust to variation in models. A reviewer of the paper pointed out that such a
finding for trust has been published before in the economic literature (Roth, 2009).
Furthermore, one small experimental study found evidence that belief in homo
economicus (economic rationality) has negative effects on trust (Xin & Liu, 2013).
In addition, economic growth has been observed to be associated with decreases
in trust, suggesting they could be causally related (Xin & Xin, 2017; Zhang & Xin,
2019). The topic needs more research.
Likewise, when we extended our study to 6 major economic outcomes
median income, median household income, GNI per capita, GDP per capita, as
well as economic growth versions of the latter two we find that national IQ
outperforms time preference on a similar scale as for the non-economic
outcomes. Thus, our findings run directly against decades of theorizing and
research in macroeconomics about the importance of time preference for
economic growth (Dohmen et al., 2015, 2018). Our findings, however, are in line
with findings about economic outcomes and national intelligence (Christainsen,
2013, 2020; Hanushek & Woessmann, 2012; Jones, 2011; Lynn & Vanhanen,
2001, 2002; Meisenberg, 2012).
Despite these strengths, the present study has some notable limitations.
First, as we noted earlier, the coverage of countries is not wholly satisfactory.
While the national IQ datasets provide data for every country, in fact a large
number of these values are estimated based on neighboring countries by spatial
nearest neighbor imputation (see Kirkegaard, 2015). This is a valid method of
data imputation when data are strongly spatially autocorrelated, as the country-
level data are. But even for the countries where data exist, they are not always
satisfactory. Since the publication of national IQ datasets, critics have repeatedly
made the point that many samples are small, unrepresentative, or very old
(Barnett & Williams, 2004; Ebbesen, 2020; Hunt & Sternberg, 2006; Wicherts,
Dolan & van der Maas, 2010).
These points are true to some extent, though there is nothing one can do
about it apart from collecting more data and refining aggregation methods. There
is a trade-off between data quality and coverage. The stricter the requirement for
data quality, the fewer countries will be covered, and the more model bias from
an unrepresentative sample, as well as less model precision owing to the smaller
sample size. The policy of the various IQ researchers has generally been to
maximize the data coverage, and then let other researchers deal with the data
quality problems as they wish.9 The up to date national IQ database by David
Becker ( provides a sample quality index for each sample
and each country, thus enabling others to vary the inclusion criteria or use quality-
based case weights to see how this affects their results. (The Rieger et al. time
preference dataset also provides quality-based case weights, though these were
not used here.) Critics have in general not done such work, nor have they
collected any additional data to test the robustness of the national IQ estimates.
Thus, they have done essentially nothing to improve the measurement of national
IQs, merely muddied the waters. This behavior suggests lack of genuine interest
in the scientific question, and thus some other motivation must explain why they
engage with this work, but only “from a distance” (Snyderman & Rothman, 1988).
There is even poorer data coverage in terms of countries for non-cognitive
traits than for intelligence. Even the great effort by Falk et al. only covered 76
countries, the lesser effort by Wang et al. covered 53, and the combined dataset
of Rieger et al. has 114 countries covered. Our analyses of the standard
deviations of the chief variables showed that these induce some level of range
restriction in the models, which there is unfortunately nothing we could do to
avoid. To the extent that our estimates of time preference and, especially, national
intelligence were error-prone, this would have the effect of weakening our results.
As a matter of fact, the sampling (in terms of sample size and representativeness)
for measurement of national IQs is in general worse than that for the non-cognitive
data, and so we would expect the national IQs to have more random error. Our
results that national IQs are nonetheless 3 to 4 times more predictive than time
preference is of note, and probably underestimates the true validity. This is a point
often missed by critics (Jones & Schneider, 2006).
One major problem with the non-cognitive data in particular is that these have
unknown psychometric properties across countries. In particular, there are no
published measurement invariance studies of them. It is known from other non-
cognitive data that these generally fail measurement invariance between
countries (Davidov et al., 2018; Munck, Barber & Torney-Purta, 2018; Muthén &
Asparouhov, 2018; Steenkamp & Baumgartner, 1998). Such failures may
represent the reference group effect (Heine, Buchtel & Norenzayan, 2008; Heine
9 The study by Jones and Schneider (2006) provides an illustrative case of researchers
doing their own quality control using the complete data from Lynn’s datasets. The
authors of that study decided that neighboring country-imputed data were not useful,
discarded countries with only small samples as well as countries with only data from
emigrants residing in another country (who might not be representative or who have
environmental advantages in their new country).
et al., 2002), also called shifting standards model (Biernat, 2009), translation
issues, or other issues such as extreme response bias (Meisenberg, 2015;
Meisenberg & Williams, 2008).
Worse, sometimes datasets gathered by different authors that purport to
measure the same construct do not correlate better than chance, but are still
employed. As a case in point, Hanushek et al. (2020) showed that the GPS
measure of time preference correlates -.06 with the estimate of long-term
orientation in the World Values Survey, and .247 with that of Hofstede. Similarly,
The GPS measure of risk-taking correlated at .239 with the same measure in the
World Values Survey, and -.302 with uncertainty avoidance in the Hofstede
dataset. Even for the two modern measures of time preference used in this study,
their country-level correlation was only about .60 (cf. Table 2). Clearly, things are
far less than satisfactory with regards to convergent validity for non-cognitive
While the situation is not perfect for the cognitive data, there is a diverse set
of cognitive indicators all of which are very strongly related, sometimes estimated
to be correlated at 1.0 when adjusted for measurement error (Lynn & Becker,
2019; Lynn & Meisenberg, 2010; Lynn & Mikk, 2007; Rindermann, 2007, 2018).
Kirkegaard (2019a) even showed that one can extract a general mental gaming
ability score from a diverse set of 12 mental games, and this correlates .87 with
Lynn’s national IQs. Furthermore, there are many within-country studies of
measurement invariance which generally support this for major population
groups, with minimal bias in reported effect sizes (Beaujean & McGlaughlin, 2014;
Frisby & Beaujean, 2015; Lasker et al., 2019). There are some studies of
measurement invariance of national IQ estimates, finding bias (Wu, Li & Zumbo,
2007) or not (Bowden, Saklofske & Weiss, 2011). Clearly, more work is needed
on this topic, and some is underway (Lasker et al., 2019). It should be noted that
when measurement invariance fails, it does not necessarily mean that the
observed international gaps are too large. They may be too small, or about the
same. The problem is that when measurement invariance fails, the measurement
properties seem to not work the same across samples, and thus the observed
results are hard to compare at the construct level. A high priority for future work
on national intelligence, and other psychological variables, is applying state of the
art measurement methods to find ways to validly compare traits across countries
which can then be used in studies such as the present.
With regards to causality, our study is merely cross-sectional and focused on
the robustness of associations. This is a weak test of causality, as it does not rule
out reverse causation, nor confounding by predictors not considered here (‘lurking
variables’). There are, however, published works on national intelligence (under
various labels) and economic outcomes, finding that reverse causality scenarios
are not supported by the data (Christainsen, 2013, 2020; Jones & Schneider,
2010; Weede & Kämpf, 2002). As with any observational study, it is possible that
we missed some very important covariate that explains away the predictive
validity of national IQ. There are not high hopes for such a variable. One prior
study used Bayesian model averaging to do an advanced ‘horse race’ between a
large collection of over 20 predictors thought to be important to explain cross-
country outcomes (e.g., origin of the legal system, Catholic % of population,
access to water) (Jones & Schneider, 2006). This study found that national IQs
were the most robust predictor in the dataset examined in the sense that among
the thousands of models examined, the best models almost invariably contained
the national IQ variable. It would be informative to re-do this approach but with a
varied set of outcome variables such as the 51+6 indicators of well-being used in
the present study, as well as other variables that have since emerged in the
scientific literature.
The causal interpretation of the present study is supported also by the results
from over 100 years of research at the individual level. There is pervasive
evidence of intelligence’s causal impact on life outcomes (Gottfredson, 1997,
2002; Herrnstein & Murray, 1994; Strenze, 2007). Thus, it does not take a genius
to realize that just as smarter people end up in better positions in life, so do
smarter countries.
Supplementary material
Supplementary materials can be found at:
R notebook:
Study materials:
Angrist, N., Djankov, S., Goldberg, P. & Patrinos, H.A. (2019). Measuring human
capital. SSRN Scholarly Paper ID 3339416, Social Science Research Network.
Barnett, S.M. & Williams, W. (2004). National intelligence and the emperor’s new
clothes. Contemporary Psychology 49(4): 389-396.
Bavel, J.J.V. & Cunningham, W.A. (2017). Scientific replication in the study of social
animals. In: J. Aronson & E. Aronson (eds.), The Social Animal, 12th edition. New York:
Beaujean, A.A. & McGlaughlin, S.M. (2014). Invariance in the Reynolds Intellectual
Assessment Scales for Black and White referred students. Psychological Assessment
26: 1394-1399.
Benjamin, D.J., Berger, J.O., Johannesson, M., Nosek, B.A., Wagenmakers, E.-J., Berk,
R., Bollen, K.A., Brembs, B., … & Johnson, V.E. (2018). Redefine statistical
significance. Nature Human Behaviour 2(1): 6-10.
Beran, M.J. & Hopkins, W.D. (2018). Self-control in chimpanzees relates to general
intelligence. Current Biology 28: 574-579.
Biernat, M. (2009). Stereotypes and shifting standards. In: T.D. Nelson (ed.), Handbook
of Prejudice, Stereotyping, and Discrimination, pp. 137-152. New York: Psychology
Bowden, S.C., Saklofske, D.H. & Weiss, L.G. (2011). Augmenting the core battery with
supplementary subtests: Wechsler Adult Intelligence Scale-IV measurement invariance
across the United States and Canada. Assessment 18: 133-140.
Carl, N. (2014). Does intelligence explain the association between generalized trust and
economic development? Intelligence 47: 83-92.
Christainsen, G.B. (2013). IQ and the wealth of nations: How much reverse causality?
Intelligence 41: 688-698.
Christainsen, G.B. (2020). Rushton, Jensen, and the wealth of nations: Biogeography
and public policy as determinants of economic growth. Mankind Quarterly 60: 458-486.
Christensen, A.P. & Golino, H. (2020). On the equivalency of factor and network
Davidov, E., Dülmer, H., Cieciuch, J., Kuntz, A., Seddig, D. & Schmidt, P. (2018).
Explaining measurement nonequivalence using multilevel structural equation modeling:
The case of attitudes toward citizenship rights. Sociological Methods & Research 47:
Dohmen, T., Enke, B., Falk, A., Huffman, D. & Sunde, U. (2015). Patience and the
wealth of nations. Unpublished Manuscript, University of Bonn.
Dohmen, T., Enke, B., Falk, A., Huffman, D. & Sunde, U. (2018). Patience and
comparative development. Working Paper.
Dutton, E., van der Linden, D. & Madison, G. (2019). Why do high IQ societies differ in
intellectual achievement? The role of schizophrenia and left-handedness in per capita
scientific publications and Nobel Prizes. Journal of Creative Behavior.
Ebbesen, C.L. (2020). Flawed estimates of cognitive ability in Clark et al. Psychological
Science, 2020 [Preprint]. PsyArXiv.
Falk, A., Becker, A., Dohmen, T.J., Huffman, D. & Sunde, U. (2016). The preference
survey module: A validated instrument for measuring risk, time, and social preferences.
Falk, A., Becker, A., Dohmen, T., Enke, B., Huffman, D. & Sunde, U. (2018). Global
evidence on economic preferences. Quarterly Journal of Economics 133: 1645-1692.
Fedderke, J.W., Klitgaard, R.E., MacMurray, J.P. & Napolioni, V. (2014). Diagnosing
deep roots of development: Genetic, disease and environmental factors. ERSA
Working Paper 465.
Frisby, C.L. & Beaujean, A.A. (2015). Testing Spearman’s hypotheses using a bi-factor
model with WAIS-IV/WMS-IV standardization data. Intelligence 51: 79-97.
Fuerst, J. & Kirkegaard, E.O.W. (2016). Admixture in the Americas: Regional and
national differences. Mankind Quarterly 56: 255-373.
Galor, O. & Özak, Ö. (2016). The agricultural origins of time preference. American
Economic Review 106: 3064-3103.
Gottfredson, L.S. (1997). Why g matters: The complexity of everyday life. Intelligence
24: 79-132.
Gottfredson, L.S. (2002). Where and why g matters: Not a mystery. Human
Performance 15: 25-46.
Grigoriev, A.A. & Lapteva, E.M. (2018). Intellectual competitiveness of a country: The
problem of national IQ mediation. Experimental Psychology (Russia) 11(3): 152-162.
Hafer, R.W. (2016). Cross-country evidence on the link between IQ and financial
development. Intelligence 55: 7-13.
Hanushek, E.A., Kinne, L., Lergetporer, P. & Woessmann, L. (2020). Culture and
student achievement: The intertwined roles of patience and risk-takings. NBER Working
Paper No. 27484.
Hanushek, E.A. & Woessmann, L. (2012). Do better schools lead to more growth?
Cognitive skills, economic outcomes, and causation. Journal of Economic Growth 17(4):
Heine, S.J., Buchtel, E.E. & Norenzayan, A. (2008). What do cross-national
comparisons of personality traits tell us? The case of conscientiousness. Psychological
Science 19: 309-313.
Heine, S.J., Lehman, D.R., Peng, K. & Greenholtz, J. (2002). What’s wrong with cross-
cultural comparisons of subjective Likert scales? The reference-group effect. Journal of
Personality and Social Psychology 82: 903-918.
Herrnstein, R.J. & Murray, C.A. (1994). The Bell Curve: Intelligence and Class Structure
in American Life. New York: Simon & Schuster.
Hillemann, F., Bugnyar, T., Kotrschal, K. & Wascher, C.A. (2014). Waiting for better, not
for more: Corvids respond to quality in two delay maintenance tasks. Animal
Behaviour 90: 1-10.
Hunt, E. & Sternberg, R.J. (2006). Sorry, wrong numbers: An analysis of a study of a
correlation between skin color and IQ. Intelligence 34: 131-137.
Jensen, A.R. (1980). Bias in Mental Testing. New York: Free Press.
Jones, G. (2011). National IQ and national productivity: The hive mind across Asia.
Asian Development Review 28: 51-71.
Jones, G. (2012). Will the intelligent inherit the earth? IQ and time preference in the
global economy. Working Paper George Mason University.
Jones, G. & Potrafke, N. (2014). Human capital and national institutional quality: Are
TIMSS, PISA, and national average IQ robust predictors? Intelligence 46: 148-155.
Jones, G. & Schneider, W.J. (2006). Intelligence, human capital, and economic growth:
A Bayesian averaging of classical estimates (BACE) approach. Journal of Economic
Growth 11: 71-93.
Jones, G. & Schneider, W.J. (2010). IQ in the production function: Evidence from
immigrant earnings. Economic Inquiry 48: 743-755.
Kirkegaard, E.O.W. (2014). The international general socioeconomic factor: Factor
analyzing international rankings. Open Differential Psychology.
Kirkegaard, E.O.W. (2015). Some methods for measuring and correcting for spatial
autocorrelation. The Winnower.
Kirkegaard, E.O.W. (2016). Inequality across US counties: An S factor analysis. Open
Quantitative Sociology & Political Science.
Kirkegaard, E.O.W. (2019a). Is national mental sport ability a sign of intelligence? An
analysis of the top players of 12 mental sports. Mankind Quarterly 59: 296-311.
Kirkegaard, E.O.W. (2019b). Solid numbers, missed opportunities: Review of The
intelligence of nations. Evolutionary Behavioral Sciences.
Kirkegaard, E.O.W. (2020, August 28). Macroeconomics and intelligence: A collection.
Clear Language, Clear Mind.
Koepke, A.E., Gray, S.L. & Pepperberg, I.M. (2015). Delayed gratification: A grey parrot
(Psittacus erithacus) will wait for a better reward. Journal of Comparative Psychology
129: 339-346.
Landes, D. (1998). The Wealth and Poverty of Nations. London: Abacus.
Lasker, J., Pesta, B.J., Fuerst, J.G.R. & Kirkegaard, E.O.W. (2019). Ancestry and IQ:
The effects of ancestry on cognitive ability in African and European-Americans. Psych
1(1): 431-459.
León, F.R. (2018). Diminished UV radiation enhances national cognitive ability, wealth,
and institutions through health and education. Personality and Individual Differences
120: 52-57.
Levine, R.V. & Norenzayan, A. (1999). The pace of life in 31 countries. Journal of
Cross-Cultural Psychology 30: 178-205.
Lv, Z. & Xu, T. (2016). The impact of national IQ on longevity: New evidence from
quantile regression. Personality and Individual Differences 101: 282-287.
Lynn, R. & Becker, D. (2019). The Intelligence of Nations. London: Ulster Institute for
Social Research.
Lynn, R. & Meisenberg, G. (2010). National IQs calculated and validated for 108
nations. Intelligence 38: 353-360.
Lynn, R. & Mikk, J. (2007). National differences in intelligence and educational
attainment. Intelligence 35: 115-121.
Lynn, R. & Vanhanen, T. (2001). National IQ and economic development: A study of
eighty-one nations. Mankind Quarterly 41: 415-435.
Lynn, R. & Vanhanen, T. (2002). IQ and the Wealth of Nations. Praeger.
Lynn, R. & Vanhanen, T. (2012). Intelligence: A Unifying Construct for the Social
Sciences. London: Ulster Institute for Social Research.
Meisenberg, G. (2012). National IQ and economic outcomes. Personality and Individual
Differences 53: 103-107.
Meisenberg, G. (2015). Do we have valid country-level measures of personality?
Mankind Quarterly 55: 360-382.
Meisenberg, G. & Williams, A. (2008). Are acquiescent and extreme response styles
related to low intelligence and education? Personality and Individual Differences 44:
Minkov, M., Welzel, C. & Bond, M.H. (2016). The impact of genes, geography, and
educational opportunities on national cognitive achievement. Learning and Individual
Differences 47: 236-243.
Mischel, W., Ebbesen, E.B. & Raskoff Zeiss, A. (1972). Cognitive and attentional
mechanisms in delay of gratification. Journal of Personality and Social Psychology 21:
Mischel, W., Shoda, Y. & Rodriguez, M.I. (1989). Delay of gratification in children.
Science 244: 933-938.
Munck, I., Barber, C. & Torney-Purta, J. (2018). Measurement invariance in comparing
attitudes toward immigrants among youth across Europe in 1999 and 2009: The
alignment method applied to IEA CIVED and ICCS. Sociological Methods & Research
47: 687-728.
Murphy, R. & Nowrasteh, A. (2016). The deep roots of economic development in the
U.S. states. SSRN Scholarly Paper ID 2822137.
Muthén, B. & Asparouhov, T. (2018). Recent methods for the study of measurement
invariance with many groups: Alignment and random effects. Sociological Methods &
Research 47: 637-664.
Nikolaev, B. & Salahodjaev, R. (2016). The role of intelligence in the distribution of
national happiness. Intelligence 56: 38-45.
O’Boyle, E.H., Humphrey, R.H., Pollack, J.M., Hawver, T.H. & Story, P.A. (2011). The
relation between emotional intelligence and job performance: A meta-analysis. Journal
of Organizational Behavior 32: 788-818.
Potrafke, N. (2019). Risk aversion, patience and intelligence: Evidence based on macro
data. Economics Letters 178: 116-120.
Rieger, M.O., Wang, M. & Hens, T. (2020). Universal time preference. Quantitative
Finance and Risk Analysis, Working Paper Series No20-07. https://www.uni-
Rindermann, H. (2007). The g-factor of international cognitive ability comparisons: The
homogeneity of results in PISA, TIMSS, PIRLS and IQ-tests across nations. European
Journal of Personality 21: 667-706.
Rindermann, H. (2018). Cognitive Capitalism: Human Capital and the Wellbeing of
Nations. University Printing House.
Rindermann, H. & Becker, D. (2018). FLynn-effect and economic growth: Do national
increases in intelligence lead to increases in GDP? Intelligence 69: 87-93.
Rindermann, H., Kodila-Tedika, O. & Christainsen, G. (2015). Cognitive capital, good
governance, and the wealth of nations. Intelligence 51: 98-108.
Roth, F. (2009). Does too much trust hamper economic growth? Kyklos 62: 103-128.
Salgado, J.F. (1995). Situational specificity and within-setting validity variability. Journal
of Occupational and Organizational Psychology 68: 123-132.
Schmidt, F.L. & Hunter, J.E. (1984). A within setting empirical test of the situational
specificity hypothesis in personnel selection. Personnel Psychology 37: 317-326.
Schmidt, F.L., Oh, I. & Shaffer, J. (2016). The validity and utility of selection methods in
personnel psychology: Practical and theoretical implications of 100 Years... Retrieved
July, 3, 2018.
Shamosh, N.A. & Gray, J.R. (2008). Delay discounting and intelligence: A meta-
analysis. Intelligence 36: 289-305.
Snyderman, M. & Rothman, S. (1988). The IQ Controversy, the Media and Public
Policy. Transaction Publishers.
Sobel, D. (2010). Longitude: The True Story of a Lone Genius Who Solved the Greatest
Scientific Problem of His Time, 1st edition. Bloomsbury.
Spolaore, E. & Wacziarg, R. (2013). How deep are the roots of economic development?
Journal of Economic Literature 51: 325-369.
Steenkamp, J.E.M. & Baumgartner, H. (1998). Assessing measurement invariance in
crossnational consumer research. Journal of Consumer Research 25: 78-107. JSTOR.
Strenze, T. (2007). Intelligence and socioeconomic success: A meta-analytic review of
longitudinal research. Intelligence 35: 401-426.
Suter, W.N. & Suter, P.M. (2019). Understanding replication: Trust but verify. Home
Health Care Management & Practice 31(4): 207-212.
Thies, C.F. (2019). GMAT scores as a proxy for national IQ. Journal of Organizational
Psychology 19(1): Article 1.
Wang, M., Rieger, M.O. & Hens, T. (2016). How time preferences differ: Evidence from
53 countries. Journal of Economic Psychology 52: 115-135.
Watts, T.W., Duncan, G.J. & Quan, H. (2018). Revisiting the marshmallow test: A
conceptual replication investigating links between early delay of gratification and later
outcomes. Psychological Science 29: 1159-1177.
Weede, E. & Kämpf, S. (2002). The impact of intelligence and institutional
improvements on economic growth. Kyklos 55: 361-380.
Wicherts, J.M., Dolan, C.V. & van der Maas, H.L.J. (2010). A systematic literature
review of the average IQ of sub-Saharan Africans. Intelligence 38: 1-20.
World Bank (2018). World Development Report 2019: The Changing Nature of Work.
Wu, A.D., Li, Z. & Zumbo, B.D. (2007). Decoding the meaning of factorial invariance and
updating the practice of multi-group confirmatory factor analysis&58; A demonstration
with TIMSS data. Practical Assessment 12(3): 1-26.
Xin, Z. & Liu, G. (2013). Homo economicus belief inhibits trust. PLoS ONE 8(10):
Xin, Z. & Xin, S. (2017). Marketization process predicts trust decline in China. Journal of
Economic Psychology 62: 120-129.
Zhang, Y. & Xin, Z. (2019). Rule comes first: The influences of market attributes on
interpersonal trust in the marketization process. Journal of Social Issues 75: 286-313.
Zissman, C. & Ganzach, Y. (2020). In a representative sample grit has a negligible
effect on educational and economic success compared to intelligence. Social
Psychological and Personality Science, 1948550620920531.
... Potential causes of externalities arising from IQ include its association with free-market opinions (Carl, 2014a(Carl, , 2015, with greater knowledge of economics (Caplan & Miller, 2010), with lower time preference and more saving (Jones & Podemska-Midluch, 2010;Kirkegaard & Karlin, 2020;Shamosh & Gray, 2008;Yeh et al., 2021), with higher levels of social trust (Carl, 2014b;Carl & Billari, 2014), with cooperation in public goods games (Al-Ubaydli et al., 2016;Putterman et al., 2012) and the prisoner's dilemma (Proto et al., 2019), national IQ's association with institutional quality (Jones & Potrafke, 2014;Kanyama, 2014;Potrafke, 2012) and the prevalence of O-ring production functions (Jones, 2013). For an overview of the mechanisms by which IQ could create externalities see Jones (2016) and Anomaly and Jones (2021). ...
... Time preference may influence economic growth through higher savings and investment (Jones, 2010(Jones, , 2012. Unsurprisingly, national IQ correlates with savings rates (Jones, 2010) and time preference measures (Kirkegaard & Karlin, 2020) at the national level. When Karlin and Kirkegaard tested national IQ and time preference as predictors of national welfare, they found time preference was statistically insignificant when IQ was included. ...
... UV radiation, time preference, social trust, and economic freedom did not have a higher PIP than prior (Figures 4 & 5). This replicates the findings of Carl (2014) and Kirkegaard and Karlin (2020) who respectively tested whether social trust and time preference could explain national IQ's relationship with economic growth. The failure of UV radiation and latitude to robustly predict economic growth suggests that IQ's relationship with growth is not due to spatial autocorrelation. ...
Full-text available
Since Lynn and Vanhanen's book IQ and the Wealth of Nations (2002), many publications have evidenced a relationship between national IQ and national prosperity. The strongest statistical case for this lies in Jones and Schneider's (2006) use of Bayesian model averaging to run thousands of regressions on GDP growth (1960-1996), using different combinations of explanatory variables. This generated a weighted average over many regressions to create estimates robust to the problem of model uncertainty. We replicate and extend Jones and Schneider's work with many new robustness tests, including new variables, different time periods, different priors and different estimates of average national intelligence. We find national IQ to be the "best predictor" of economic growth, with a higher average coefficient and average posterior inclusion probability than all other tested variables (over 67) in every test run. Our best estimates find a one point increase in IQ is associated with a 7.8% increase in GDP per capita, above Jones and Schneider's estimate of 6.1%. We tested the causality of national IQs using three different instrumental variables: cranial capacity, ancestry-adjusted UV radiation, and 19 th-century numeracy scores. We found little evidence for reverse causation, with only ancestry-adjusted UV radiation passing the Wu-Hausman test (p < .05) when the logarithm of GDP per capita in 1960 was used as the only control variable.
Full-text available
This paper offers a review of some of the empirical literature on economic growth and discusses its recent evolution in light of developments in intelligence research and genomics. The paper also undertakes the first regression analysis of economic growth to use the most up-to-date version (VI.3.2) of David Becker's data set of international IQ scores. The analysis concerns the growth of 94 countries from 1995-2016. The new regression analysis replicates the results of Jones and Schneider (2006) in finding IQ to have a robust impact on economic growth. Political and economic institutions are represented in the regressions via a country's "degree of capitalism" (aka "economic freedom"), which is found to have an impact that is positive and statistically significant. A change from communism to a market economy does much to increase growth, but the paper finds diminishing returns to free markets. Countries whose people are mostly of sub-Saharan African descent have low average IQ scores, but the paper finds that other factors also have lessened economic growth not only in Africa, but in Haiti and Jamaica as well. Rushton and Jensen (2005, 2010) put forth the hypothesis that average IQ differences across ethnic groups are 50% due to genetic differences, and 50% due to differences in natural and social environments. Applied to international IQ scores, the paper finds the hypothesis to be very reasonable.
Full-text available
Research at the individual level shows strong positive relationships between performance in video games and on intelligence tests. Together with evidence of above average IQs of players of traditional mental sports such as chess, this suggests that national IQs should be strongly related to national performance on mental sports. To investigate this, lists of top players for 12 different electronic sports (e-sports) and traditional mental sports were collected from a variety of sources (total n = 36k). Using a log count approach to control for population size, national cognitive ability/IQ was found to be a predictor (p<.05) of the relative representation of countries among the top players for every game except Go. When an overall mental sports score was calculated using a factor analytic approach, the factor scores correlated r = .79 with Lynn and Vanhanen's (2012) published national IQs. The pattern was somewhat nonlinear such that national IQs below 85 seemed to have no relationship. The games that related most strongly with the general factor of mental sport ability also correlated more strongly with national IQs (r = .94). The relationship was fairly robust to controls for geographical region (coefficient 74% of the original in chosen model specification).
This paper studies the relationship between patience and comparative development through a combination of reduced-form analyses and model estimations. Based on a globally representative dataset on time preference in 76 countries, we document two sets of stylized facts. First, patience is strongly correlated with per capita income and the accumulation of physical capital, human capital and productivity. These correlations hold across countries, subnational regions, and individuals. Second, the magnitude of the patience elasticity strongly increases in the level of aggregation. To provide an interpretive lens for these patterns, we analyze an OLG model in which savings and education decisions are endogenous to patience, aggregate production is characterized by capital-skill complementarities, and productivity implicitly depends on patience through a human capital externality. In our model estimations, general equilibrium effects alone account for a non-trivial share of the observed amplification effects, and an extension to human capital externalities can quantitatively match the empirical evidence.
We compare the relative contribution of grit and intelligence to educational and job-market success in a representative sample of the American population. We find that, in terms of Δ R ² , intelligence contributes 48–90 times more than grit to educational success and 13 times more to job-market success. Conscientiousness also contributes to success more than grit but only twice as much. We show that the reason our results differ from those of previous studies which showed that grit has a stronger effect on success is that these previous studies used nonrepresentative samples that were range restricted on intelligence. Our findings suggest that although grit has some effect on success, it is negligible compared to intelligence and perhaps also to other traditional predictors of success.
In a recent paper, Clark et al. (2020) analyze the relationship between socio-economic measures and estimates of average cognitive ability (’IQ’) in 140 countries. According to the data presented by the authors, several African, South Asian and Central American countries have an average IQ below 50 (i.e. intellectually disabled, according to DSM diagnostic criteria). Moreover, according to the data presented by the authors, the average cognitive ability of adults in African nations is ∼1.6 standard deviations below the cognitive ability of European adults. These notions are incompatible with psychological science and all conclusions drawn from these data are invalid.
Previous research has attempted to understand why countries with relatively favorable conditions and high estimated average IQs (such as Finland and Japan) have a relatively low per capita number of scientific Nobel prizes. In the present study, we examine whether there is a relationship between national schizophrenia and left‐handedness prevalence, on the one hand, and per capita scientific and literary achievement, on the other hand, in countries with IQ estimates of at least 90. We found that per capita science and literature Nobel prizes and scientific publications are strongly negatively associated with schizophrenia and strongly positively correlated with left‐handedness. There also was a very pronounced negative correlation between schizophrenia rate and left‐handedness rate. These results suggest that genius can be regarded as a combination of very high IQ, aspects of high‐functioning autism (specifically low empathy) plus relatively low impulse control, consistent with observations of intellectually outstanding individuals, and the fact that schizophrenia appears to constitute the opposite pole of these aspects of autism spectrum. We posit differences in androgen levels as a possible underlying explanation for these findings.
Replication as a pillar of science is described in the context of the replication crisis that first struck psychology but spread quickly to other science-based fields. Empirical evidence suggests that the crisis is real but not well understood. We explain why replication often fails in science and how research in home health can be strengthened by a greater understanding of the value of replication and current thinking about replication success or failure decisions. We conclude with a call for replication in home health nursing that couples original research and replication in the same report.
Using the new macro data on risk aversion and patience by Falk et al. (2018), I show that risk aversion and patience are related to intelligence: high-IQ populations are more patient and more risk averse than low-IQ populations. The correlation between patience and intelligence corroborates previous results based on micro data. Intelligent people tend to be patient because they have long time horizons. The correlation between risk aversion and intelligence supports new micro data studies based on dynamically optimized sequential experimentation (Chapman et al. 2018).