ArticlePDF Available

Danish Exceptionalism: Explaining the Unique Increase in Social Trust Over the Past 30 Years

Authors:

Abstract and Figures

Given the positive societal consequences of social trust and the exceptional high levels of trust found in the Nordic countries, there is a growing interest in the background for Nordic trust exceptionalism. Reflecting the debate about the roots of trust, the question is whether this exceptionalism has been a permanent feature of these societies, i.e., an enduring cultural trait, or primarily been shaped by more contemporary experiential forces. This article examines the roots of the present Nordic trust exceptionalism by means of analysing the development of trust in Denmark over the past three decades. The analysis shows that trust in Denmark has increased remarkably from 1979 to 2009, thus documenting that the high levels of trust found today are a relatively new phenomenon. This contradicts the cultural perspective and instead vindicates the experiential perspective on trust. Subsequent analyses based on longitudinal data at the aggregate level as well as individual-level cross-sectional and panel data indicate that the increase in trust in Denmark can be attributed to generational replacement, increasing levels of education, improved quality of state institutions, and a concomitant increase in citizens’ trust in these institutions.
Content may be subject to copyright.
Danish Exceptionalism: Explaining the Unique Increase in Social Trust over the Past 30
Years
Kim Mannemar Sønderskov & Peter Thisted Dinesen
[Preprint version, final version is forthcoming in European Sociological Review]
Abstract
Given the positive societal consequences of social trust and the exceptional high levels of trust
found in the Nordic countries there is a growing interest in the background for Nordic trust
exceptionalism. Reflecting the debate about the roots of trust, the question is whether this
exceptionalism has been a permanent feature of these societies, i.e. an enduring cultural trait, or
primarily been shaped by more contemporaneous experiential forces. This paper examines the roots
of the present Nordic trust exceptionalism by means of analyzing the development of trust in
Denmark over the last three decades. The analysis shows that trust in Denmark has increased
remarkably from 1979 to 2009 thus documenting that the very high levels of trust found today are a
relatively new phenomenon. This result contradicts the cultural perspective and instead vindicates
the experiential perspective on trust. Subsequent analyses based on longitudinal data at the
aggregate level as well as individual-level cross-sectional and panel data indicate that the increase
in trust in Denmark can be attributed to generational replacement, increasing levels of education,
improved quality of state institutions, and a concomitant increase in citizens’ trust in these
institutions.
Keywords
Keywords: social trust, institutional trust, institutional quality, education, panel data, Denmark
The author listing purely reflects reversed alphabetical order of the last names as the authors contributed equally to the
article.
1
Introduction
Social trust is the belief that most people with whom one could potentially interact will behave in a
trustworthy manner. A person’s level of social trust thus reflects her standard estimate of the
trustworthiness of the generalized, anonymous other (Coleman, 1990: 104; Robinson & Jackson,
2001).1 Social trust is an important resource in modern societies, where citizens constantly interact
with anonymous others whom they have no specific knowledge about. In many everyday
interactions, e.g., shopping or driving in traffic, as well as more abstract and complex situations
such as forming attitudes about political issues or deciding whether to pay taxes, citizens’ behavior
depend on their expectations about the behavior of other citizens. In such situations involving
anonymous others, social trust serves as an important heuristic for behavior with significant positive
consequences for society as a whole (Stolle, 2001; Sønderskov, 2008). This is supported
empirically in a number of studies, which show that individuals with higher levels of trust
contribute more to the provision of public goods through, e.g., tax paying, pro-environmental
behavior, volunteering, and immunization (Gächter et al., 2004; Rönnerstrand, 2013; Scholz &
Lubell, 1998; Sønderskov, 2011a; Uslaner, 2002). Thus, there is little doubt that social trust is
valuable resource for modern societies.
Because of its positive societal consequences, it is important to understand why some countries are
characterized by high levels of social trust and others are not. In this regard, the Nordic countries –
Denmark, Sweden, Norway, Finland and Iceland – have attracted massive attention because they
are outliers placed at the very top of the trust distribution2 (see, e.g., Bjørnskov, 2006, 2008) leading
scholars to speak of “Nordic exceptionalism” (Delhey & Newton, 2005). This naturally begs the
question about the roots of this exceptionalism and, ultimately, whether other countries can learn
from the Nordic countries with regard to building up trust. On a theoretical level this taps into a
central debate about whether social trust is a long-standing cultural trait (Putnam, 1993; Uslaner,
2002), or a more manipulable characteristic that can be generated in the shorter term by
contemporary experiences (Dinesen, 2012b; 2013; Glanville & Paxton, 2007). In this regard, the
Nordic countries are often seen as hallmarks of the cultural account because they share a number of
deep historical structures known to correlate positively with trust, e.g. Protestantism and low ethnic
diversity (Delhey & Newton, 2005). On account of its historical correlates, the current elevated
levels of trust are taken as an indication that high trust is a permanent feature of the Nordic culture.
However, to show that the Nordic trust exceptionalism is indeed culturally founded, a much more
direct test is to look at trust levels over time. If the present high levels of trust are stable back in
2
time, this is a strong indication of a long-standing Nordic trust culture and thus the cultural account
of the roots of trust. In contrast, if trust changes over time, it substantiates theories emphasizing
how contemporary experiences influence trust, in turn leading to the questions of what these factors
may be.
This paper examines the roots of present Nordic exceptionalism by means of analyzing the
development of trust in Denmark – one of the global high-scorers in trust (Bjørnskov, 2006; 2008) –
over the last three decades. Using aggregate data our analysis shows that trust in Denmark has
increased remarkably from 1979 to 2009 thus documenting that the very high levels of trust found
today are a relatively new phenomenon. In theoretical terms, this result contradicts the cultural
perspective and instead vindicates the experiential perspective on trust. Subsequently, we scrutinize
a number of plausible experiential explanations for the increase in trust based on a longitudinal
analysis at the aggregate level as well as individual-level cross-sectional and panel data. These
analyses indicate that the increase in trust in Denmark can be attributed to generational replacement,
increasing levels of education, improved quality of state institutions, and a concomitant increase in
citizens’ trust in these institutions.
Explanations of social trust: Cultural and experiential perspectives
As just noted, a central distinction within theories of the roots of trust is that between cultural and
experiential explanations. The cultural perspective emphasizes how trust is passed on from parents
to children through early life socialization (Uslaner, 2002; 2008)3 with little subsequent role for
experiential alteration. As a consequence, the cultural perspective predicts levels of trust to be
highly stable over the individuals’ life course and over generations and therefore also at the society
level. Illustrative of the cultural perspective, Uslaner (2008) shows that the level of trust of
immigrants in the USA tends to reflect that of their grandparents’ country of origin (see also Rice &
Feldman, 1997 and Tabellini, 2008). According to this perspective, Danish (and Nordic) trust
exceptionalism, should thus be explained by a trust culture set in motion by forces dating back very
far in time. In this regard, Svendsen & Svendsen (2004) argue that the Danish co-operative
movement, which arose among peasants in the late 19th century, was built on a micro-foundation of
trust and social capital among neighboring farmers, and that the movement further expanded the
stock of society-level social capital, which in turn strongly influenced the later political and
institutional development in Denmark with positive consequences for social trust.
3
In contrast to the cultural explanation, the experiential explanation emphasizes the susceptibility of
trust to various experiences. Such experiences are typically perceived broadly and thus encompass
both more direct personal experiences (e.g. with other people in various settings), but also collective
experiences influencing all individuals in a society (e.g. income inequality or wide-spread
corruption) (Uslaner, 2002). The essence of the experiential perspective is that trust is malleable and
thus probably not temporally stable. This perspective has also been substantiated by studies of
immigrants in Western Europe, showing that destination country environments conducive to trust
exercise a marked impact on immigrants' trust (Dinesen, 2012a, 2012b, 2013; Nannestad et al.,
2014). In experiential terms, Danish trust exceptionalism, and changes in trust over time, would
then be explained by trust-promoting experiences. In this regard, level of education, ethnic and
economic diversity and the quality of state institutions are often emphasized as important
explanations (Bjørnskov, 2006; 2008; Brehm & Rahn, 1997; Delhey & Newton, 2003; 2005;
Dinesen, 2013; Dinesen & Sønderskov, 2012a; Nannestad et al., 2014; Rothstein & Stolle, 2008;
Rothstein & Uslaner 2005).
Having sketched the two main perspectives on the roots of trust above, it should be clear that they
yield different predictions about the temporal dynamics of trust. The cultural perspective predicts
temporal stability regardless of environmental developments, whereas the experiential perspective
predicts increasing/decreasing trust in the face of major societal changes. Below we examine the
leverage of the two perspectives in explaining Danish trust exceptionalism.
Testing the cultural and the experiential perspective on trust in Denmark over time
A straightforward test of the two perspectives on trust is to examine aggregate levels of trust over
time. Using data from Denmark we are thus able to understand whether the Danish trust
exceptionalism has been a stable cultural feature or instead appear to be a more recent phenomenon.
Fortunately, Danish surveys holding a standard measure of social trust exist all the way back to
1979, thus providing a reasonably long time series to examine the dynamics of trust. Table 1 plots
the estimated share of the Danish population responding that “most people can be trusted” to the
classical dichotomous question “Generally speaking, would you say that most people can be trusted
or that you cannot be careful when dealing with others?”.4 In order to contrast the Danish
experience, we have also plotted data for a number of other countries, e.g. two other Nordic trust
high-scorers, Norway and Sweden, and the US, which is known for its historical drop in trust (Clark
& Eisenstein, 2013; Putnam, 1995; 2000; Robinson & Jackson, 2001; Uslaner, 2002).5
4
Figure 1: The development of social trust in selected countries
Notes: The level of trust is the share of the population answering that “most people can be trusted” to the question,
“Generally speaking, do you that think most people can be trusted or that you cannot be too careful in dealing with
others” after removing “Don’t know” responses. Sources: World Values Survey 1981-2008 (WWS 2009) (data for
Norway, Sweden, US, GB, France); European Values Study 1981-2008 (EVS 2011) (all countries); General Social
Survey 1972-2010 (GSS 2011) (US); Politiske værdier i Danmark 1979 (Danish Data Archive 1981) (Denmark); Social
ulighed i Danmark 2009 (Danish Data Archive 2011) (Denmark). When more than one survey contains data for the
same country in the same year, we report the pooled share expressing trust.
Figure 1 tells an unequivocally clear story with regard to the development of trust in Denmark over
the past 30 years. In 1979, 47% expressed that “most people can be trusted”, whereas this share had
increased to 79% in 2009, thus leading to a remarkable increase in trust of about a third of the
possible range over the 30 year period. This huge monotonic increase in trust strongly suggests that
trust is not temporally stable and that the very high levels of trust presently found in Denmark are a
relatively new phenomenon. This in turn implies that Danish trust exceptionalism is primarily
rooted in more recent experiences rather than a cultural heritage. The comparative picture also
underlines that Danish levels of trust have in fact only turned exceptional in more recent years. In
1979, Danish levels of trust were on a par with the level in the US in 1980, and going even further
back in time, 58% of Americans trusted most people in 1960 (Putnam, 1995), a substantially higher
share than the 47% in Denmark in 1979. Similarly, the level of trust in Denmark was only slightly
higher than that in Great Britain in 1981, and in the beginning of the eighties Denmark was lagging
.2 .4 .6 .8
Social trust
1980 1990 2000 2010
Year
Norway Denmark Great Britain
Sweden United States France
5
somewhat behind its Scandinavian counterparts. That said, Danish levels of trust in the late
seventies was fairly high by both that time’s as well as present day standards (Delhey & Newton,
2005; Bjørnskov, 2008), so we cannot rule out whether this baseline has cultural roots. However,
the emergence of Danish trust exceptionalism from 1979 onwards clearly indicates the dominance
of experiential forces in explaining the more recent development. With this in mind, we will now
turn to the question of which experiential factors underlies this development.
Explanations of the Danish increase in trust: cohort and period explanations
A central distinction within the experiential perspective relates to whether experiences at certain
points in time have different effects on trust. One line of thought emphasizes how trust is
continuously formed throughout life (Dinesen, 2012b, 2013), while another view argues that
experiences during childhood and adolescence often leave a lasting mark on trust and other attitudes
(Robinson & Jackson, 2001). The latter perspective thus shares the idea of the importance of early
life factors in shaping subsequent levels of trust with the cultural perspective (Uslaner, 2002, ch. 4),
but does not preclude the importance of experiences later in life, that is, early life experiences can
be seen as a baseline from which the influence of subsequent experiences depart. The distinction
between the two experiential perspectives parallels the discussion within the literature on attitude
formation regarding the “impressionable years” hypothesis versus the “lifelong openness to
experience” hypothesis (Sears & Levy, 2003). Similarly, it also relates to the literature on value
change distinguishing between cohort and period explanations of change (Glenn, 2005; Robinson &
Jackson, 2001).6 Cohort explanations focus on how people born around the same time are subject to
the same environment and circumstances (and thus have common experiences), and therefore tend
to have similar values over their life courses. While the common experiences shaping values may in
principle occur throughout life, it is often assumed – in accordance with the “impressionable years”
hypothesis – that experiences in late adolescence and early adulthood play an especially important
role in shaping values. Given its focus on value homogeneity within cohorts due to common
experiences, the cohort explanation explains value change by older generations with one set of
values being replaced by younger generations with another set of values. The cohort explanation of
trust has received considerable support in Robert Putnam’s analysis of social capital in the USA, in
which he finds that the main cause of decline in indicators of social capital (including social trust) is
generational replacement of “the long civic generation” born before 1930 by younger cohorts
6
(Putnam, 2000: Ch. 8, 14-15; see also Clark & Eisenstein, 2013; Robinson & Jackson, 2001;
Schwadel & Stout, 2012).
Whereas cohort explanations emphasize the importance of early-life environment, period
explanations stress how societal factors and specific events in a given period of time affect
everyone equally independent of cohort. Hence, general societal changes such as those regarding
economic redistribution, ethnic composition, and institutional set-up are expected to influence the
values of all citizens in a similar manner. In the following we evaluate these explanations using
aggregate data over time from Denmark, and subsequently test the most promising explanations
more rigorously using individual-level data.
Figure 2: The development of social trust for 20th century Danish cohorts
Notes: Social trust is measured as in Figure 1. The data point in 1980 is the pooled share expressing trust in two surveys
(1979 and 1981). The 2009 data point is the pooled share expressing trust in surveys from 2008 and 2009. Data points
based on less than 20 respondents are omitted. Sources: European Values Study 1981-2008 (EVS 2011); Politiske
værdier i Danmark 1979 (Danish Data Archive 1981) Social ulighed i Danmark 2009 (Danish Data Archive 2011).
Cohort explanations and changes in trust
In order to assess the cohort explanation of the increase in trust in Denmark over the past 30 years,
we examine the level of trust (measured between 1979 and 2009) for birth cohorts from the first
nine decades of the 20th century in Figure 2. The trends in trust among the different cohorts provide
.4 .5 .6 .7 .8
Social trust
1980 1990 2000 2010
Year of survey
1900-1909
1910-1919
Twenties
Thirties
Forties
Fifties
Sixties
Seventies
Eighties
Birth decade
7
partial evidence in favor of the cohort explanation as older cohorts (i.e., those born in the beginning
of the 20th century) display lower levels of social trust than cohorts born later in the century. In
1980, 30 percent of the population was born before 1930, while these cohorts’ share of the
population had dropped to 10 percent in 2005.7 Hence, a part of the increase in social trust in
Denmark seems to be due to the replacement of older, less trustful cohorts with younger, more
trustful cohorts.
Having identified a cohort effect begs the question of what explains this difference in trust between
younger and older generations. In this regard, rising levels of education is a prominent candidate as
higher levels of education have systematically been shown to be associated with higher levels of
trust at the individual level (Brehm & Rahn, 1997; Delhey & Newton, 2003; Putnam, 1995;
Uslaner, 2002: Ch. 5). Given the rising levels of education in Denmark over time (see Figure 3
below), it seems plausible that the replacement of older cohorts with younger, and better educated,
cohorts can explain some of the observed cohort effect in Figure 2. Consequently, we examine
education as an explanation for the cohort effect in trust in the subsequent analyses.
While Figure 2 documents a cohort effect, it also testifies to generational replacement not being the
only explanation of the rising levels of trust in Denmark as trust rises almost in parallel for all
cohorts over the period of time examined. This indicates that social trust is not a value set in stone
from early on in life, which is immune to subsequent experiences. Instead, the parallel increase in
trust over time across cohorts indicates that general societal changes – and hence period
explanations – can account for at least part of this development. This is the question we turn to next.
Period explanations and changes in trust
To explain the rising levels of trust, we examine three of the main societal factors (and thus period
explanations) identified in the literature: economic equality, ethnic homogeneity, and institutional
quality.
A number of studies have shown that citizens in more economically equal societies also tend to
have higher levels of trust. This is explained by a stronger sense of community among citizens in
these societies and by low levels of inequality reducing the potential gains from engaging in
untrustworthy behavior (Bjørnskov, 2006; 2008; Rothstein & Uslaner, 2005; Uslaner, 2002). Ethnic
homogeneity has also been identified as a potential explanation of social trust. Lower degrees of
ethnic conflict, and greater predictability of the actions of people of similar ethnic backgrounds are
some of the factors suggested to lead to higher levels of trust in ethnically homogeneous societies
8
(Delhey & Newton, 2005; Dinesen & Sønderskov, 2012a; Knack & Keefer, 1997). Finally, quality
of government in terms of fair and efficient state institutions is a third societal explanation of social
trust. According to this view, institutional quality lays the foundation for social trust by sanctioning
treacherous behavior and thus reducing incentives for behaving in an untrustworthy manner, and by
securing equal treatment of citizens and hence reduce suspicion of others. Moreover, when the
representative of fair institutions – either bureaucrats or publicly elected politicians – behave in an
incorrupt manner, they send the signal that this type of behavior is not widespread in the population
and, hence, that most people can be trusted (Rothstein & Stolle, 2008).
To get a first hint about whether rising levels of education can explain the cohort effect observed in
Figure 2, and whether income inequality, ethnic heterogeneity, and institutional quality appear to be
plausible period explanations of increasing trust in Denmark since 1979, we have illustrated the
development in these factors in Figure 3.
Panel A in Figure 3 shows the development in the level of education in Denmark and whether
looking at average schooling or the share of people with education above primary school, there has
been a steady increase over the period examined. Hence, it seems plausible that at least some of the
cohort effect found for social trust can be explained by rising levels of education, with older
cohorts, who are both less well educated and less trustful, gradually being replaced by younger
cohorts, who are better educated and more trustful.
Panel B shows that economic inequality has generally increased over the period examined, which
points to a negative relationship between trust and economic equality. This is surprising as earlier
studies generally show a negative relationship between social trust and economic inequality
(Bjørnskov, 2006; 2008; Uslaner, 2002; Rothstein & Uslaner, 2005, but see Dinesen & Sønderskov,
2013 and Fairbrother & Martin, 2013 for recent within-country studies showing no effect of
inequality). Given that this relationship is well established in the literature, the negative correlation
between trust and equality should probably not be interpreted causally. It seems more reasonable to
conclude that social trust either is unrelated to inequality or that it has increased despite of – and not
because of – economic inequality. For this reason, we will not pursue this explanation in more detail
in the present analysis.
9
Figure 3: The development of educational attainment, economic inequality, ethnic diversity, and institutional
quality/trust in Denmark
Notes: Average years of schooling is from Barro & Lee (2010) (variable name: yr_sch); Share of the population with at
least secondary education is calculated based on data from Statistics Denmark (accounts: HFU1, KRHFU1, BEF5); The
Gini-coefficient is from Finansministeriet (2010); Share of non-Western Immigrants is from Statistics Denmark
(accounts: BEF3, KRBEF3). Institutional quality is from The Worldwide Governance Indicators project (Kaufmann et
al., 2011) and is calculated as the average of the following variables: Voice and Accountability, Government
Effectiveness, Regulatory Quality, Rule of Law and Control of Corruption. Institutional trust is the average level of
trust in a number of state institutions in the adult population on a scale from 0 to 1, where 1 signifies high trust (see the
section “Measures” in the text for more details).
Panel C shows a pattern of increasing ethnic diversity over the period examined, which, on the face
of it, suggests an unexpected positive relationship between ethnic heterogeneity and trust. However,
as for the unexpected positive correlation between economic inequality and trust, this relationship
should most likely not be given a causal interpretation. In fact, Dinesen & Sønderskov (2012a)
show that increased diversity in Danish municipalities (examined over the same period as in this
paper) has a negative impact on trust, which leads them to argue that trust might have increased
even further at the national level had it not been for increased ethnic diversity (see also Dinesen &
Average years of schooling,
adult population (left axis)
Share of adult population
with at least secondary education (right axis)
0.65 0.70 0.75 0.80
910 10 11
1980 1985 1990 1995 2000 2005 2010
Panel A: Education
Gini coefficient
0.19 0.21 0.23 0.25
1980 1985 1990 1995 2000 2005 2010
Panel B: Economic inequality
Share of non-Western immigrants
0.00 0.02 0.04 0.06 0.08
1980 1985 1990 1995 2000 2005 2010
Panel C: Ethnic diversity
Institutional trust (left axis)
Institutional quality (right axis)
1.8 1.9 2.0 2.1
0.56 0.59 0.62 0.65
1980 1985 1990 1995 2000 2005 2010
Panel D: Institutional trust and quality
10
Sønderskov, 2013). Hence, we will not examine the role of ethnic homogeneity in explaining the
rise in trust in more detail in the following.
The last panel in Figure 3 maps the development in institutional quality in Denmark and trust in
institutions among Danish citizens. Apparently, institutional quality has generally improved in the
period for which data exists although the increase is waning in the last part of the period.8
Moreover, the improvements seemingly manifest themselves in citizens’ trust in state institutions,
although with a lag. Contrary to the development in economic inequality and ethnic heterogeneity,
the improved institutional quality corresponds to our theoretical expectation. Hence, improved
quality of state institutions and a subsequent increase in institutional trust seem like a plausible
explanation for the increase in social trust based on Figure 3.
Summing up, the patterns in Figures 2 and 3 suggest that the marked increase in social trust in
Denmark can be explained by two factors: first, generational replacement in terms of higher
educated, more trustful cohorts replacing older, less well-educated cohorts and, second, improved
institutional quality and subsequent increases in institutional trust. Although both explanations
square well with the literature, and as such appear to be plausible explanations for the rising levels
of trust, we have so far only presented aggregate, bivariate evidence for the two explanations. In
order to test these explanations more rigorously, we propose four hypotheses, which we test by
means of multivariate analysis using individual-level data.
Individual-level analysis
If the increase in social trust is due to generational replacement, the older cohorts, whose share of
the population is waning, should have lower levels of trust than younger cohorts. We test this
hypothesis using data from 2008. At this point in time, it is primarily cohorts born before 1930 that
are being replaced, and we therefore hypothesize:
H1: Cohorts born before 1930 have lower levels of social trust than younger cohorts.
If the increase in social trust can be attributed to rising levels of education, we should see that the
higher educated have higher levels of trust than the lower educated. Moreover, if the potential
cohort effect is partly explained by older cohorts being lower educated than younger cohorts, we
expect to find the differences between cohorts reduced when education is held constant in the
model. Hence, we propose the following hypotheses:
H2: Higher educated people have higher levels of social trust than lower educated people.
H3: The cohort effect is reduced when education is held constant.
11
The importance of institutional quality is somewhat harder to test. Country-level measures of
institutional quality coinciding with public opinion surveys on social trust are too few to allow
testing whether improvement in institutional quality is the cause of increased social trust over time.
For that reason, we cannot assess whether the increased level of social trust in Denmark can be
attributed to changes in national-level institutional quality. However, Wang & Gordon (2011) report
a cross-country relationship between institutional quality and institutional trust, which is an
individual-level indicator of perceived institutional quality, and we therefore examine whether
institutional trust affects social trust in the Danish case. If so, this would be an indication of the
importance of the role of institutional quality in explaining the development in social trust. Hence,
we hypothesize:
H4: People with higher levels of trust in state institutions have higher levels of social trust than
people with lower levels of institutional trust.
Data
In testing hypotheses H1 to H3, we utilize the Danish part of the European Value Survey from 2008
(EVS, 2011), which is a random sample of the adult population in Denmark. The survey was
conducted by means of face-to-face interviews with a response rate of 51 percent. H4 is also tested
with these cross-sectional data. However, some authors argue in favor of reverse causality and
hence that social trust affects trust in institutions (or institutional quality) (Bjørnskov, 2006; Knack,
2002; Robbins, 2012; Uslaner, 2002: 158). Thus, a correlation between institutional trust and social
trust cannot a priori be considered support for the institutional explanation of increased levels of
social trust as using cross-sectional survey data to substantiate a causal relationship between two
(attitudinal) variables is generally problematic. For that reason we also test H4 using panel data. We
build on survey data from a panel of randomly sampled individuals who were interviewed in 1990,
1999, and 2008 in connection to the Danish part of the European Value Survey (Danish Data
Archive, 2002; 2009).9 As in most panel surveys, a number of respondents dropped out of the
survey between the waves, which means that the panel survey cannot be assumed to be
representative of the original population in 1990. Despite this problem, the panel survey provides
increased leverage for examining the causal relationship between the two types of trust at the
individual level and thus strengthens the potential evidence in favor of the institutional explanation
of the increase in social trust.
12
Specifications
H1 regarding cohort effects is tested in Model I in Table 1, which include birth cohort as well as
gender and place of birth. H2 and H3 regarding the mediating effect of education are tested by
adding level of education in Model II. The first test of H4 is reported in Model III, in which
institutional trust is added along with a number of additional control variables suggested in the
literature (see, e.g., Uslaner, 2002). Specifically we include employment, income, life satisfaction,
and self-efficacy, which are all potential mediators of the impact of birth cohort and education
analyzed in Model II. Finally we also include associational involvement – a prominent explanation
of social trust – although recent studies have made a case for reverse causality, i.e. that trusting
individual self-select into voluntary organizations (e.g. Bekkers, 2012; Stolle, 2001; Sønderskov,
2011b).10
In Models IV-VII in Table 2, we examine if a correlation between institutional trust and social trust
can be given a causal interpretation using the panel survey. Specifically, in Model IV, we examine
whether institutional trust measured in the two first waves in 1990 and 1999 can explain social trust
in 2008 when simultaneously controlling for social trust in 1990 and 1999. If institutional trust
measured in the two earlier waves of the survey influences social trust in the last wave while
controlling for previous levels of social trust, we take it as evidence in favor of a causal effect of
institutional trust on social trust because previous levels of institutional trust predicts future levels
of social trust net of previous levels of social trust. However, the analysis is restricted to a very
limited sample because only 324 respondents chose to participate in all three waves of the panel-
survey (and to answer all relevant questions). In Model V, we therefore only include one lag by
including respondents who participated in 1999 (t-1) and 2008 (t0) or, alternatively, respondents who
did not participate in 2008 but in 1990 (t-1) and 1999 (t0). This allows us to test the relationship on a
larger and more representative sample. The reverse relationship between the two forms of trust (i.e.
with social trust as the dependent variable) is examined in Models VI and VII using the same logic
applied in the model with institutional trust as the dependent variable. This serves to scrutinize
whether causality runs in the other direction or both ways. Ideally, including the lagged dependent
variable would serve to take into account the impact of all alternative factors influencing trust, but
as a considerable time lag is present between the three measurements of the two types of trust, the
viability of this strategy is somewhat questionable, and we therefore include the same control
variables as in Model III (all measured in 2008) to avoid confounding of the relationship between
the two types of trust by other extraneous factors.
13
Measures
As previously, social trust is measured by the question, “Generally speaking, do you think that most
people can be trusted or that you cannot be too careful in dealing with other people?” with the two
response options, “Most people can be trusted” and “You cannot be too careful”, of which the first
indicates social trust. Birth cohort is operationalized as the decade of birth, where respondents born
before 1930 are collapsed in a single category to avoid having too few observations in the oldest
cohorts. Education is measured in five categories: primary school, vocational education, shorter
post-secondary education (1-2 years post high school), medium-long post-secondary education (3-4
years post high school), and academic education (more than 4 years post high school). Institutional
trust is measured with an index of trust in four state institutions: the parliament, the judiciary, the
police, and the public sector. Therefore, we construct an additive index based on trust in the four
institutions. The constructed index ranges from 0 (low trust) to 1 (high trust) and indicates each
respondent’s average score on the four trust items.11 Details about the coding of the control
variables as well as descriptive statistics for all variables are found in Table A in the supplementary
material.
Analysis
Table 1 displays the results of the empirical test of hypotheses H1 to H3, which are tested using the
cross-sectional survey from 2008. Since social trust is measured dichotomously, the models are
estimated using logistic regression. Model I substantiates that generational differences between
cohorts can explain part of the increase in trust in Denmark. People born before 1930 (the reference
category) have significantly lower levels of trust than all other cohorts except for the cohort born in
the 1930’s. Moreover, it should be noted that the difference in trust between the oldest and the
younger cohorts is probably underestimated given that a larger fraction of the oldest cohort has
passed away in 2008. As social trust is positively correlated with indicators of health (see, e.g.,
Rostila, 2007), it seems likely that the individuals who live longer are generally the more trustful.
14
Table 1: Causes of social trust, cross-sectional data
Model I Model II Model III
Coef. S.E. Coef. S.E. Coef. S.E.
Birth cohort *** ***
Before 1930 Reference Reference Reference
1930’s 0.18 0.31 0.03 0.33 0.08 0.36
1940’s 0.58 0.30* 0.38 0.31 0.28 0.34
1950’s 0.80 0.30** 0.43 0.32 0.23 0.37
1960’s 1.44 0.31** 1.06 0.32** 0.79 0.39*
1970’s 1.09 0.31** 0.55 0.33 0.34 0.39
1980’s 0.77 0.32* 0.31 0.34 0.19 0.38
Gender (female) 0.09 0.13 -0.09 0.14 -0.02 0.15
Place of birth (native) 1.12 0.25** 1.29 0.27** 1.29 0.27**
Education *** ***
Primary school Reference Reference
Vocational 0.51 0.17** 0.37 0.17*
Shorter post-secondary 1.13 0.19** 0.98 0.20**
Medium-long post-secondary 1.58 0.24** 1.41 0.24**
Academic 2.29 0.35** 2.05 0.38**
Institutional trust 2.03 0.48**
Employed (yes) 0.27 0.19
Relative household income
< 25. Percentile Reference
Inter quartile range -0.01 0.19
>75. Percentile 0.06 0.29
Not revealed -0.25 0.19
Associational involvement 0.42 0.15**
Life satisfaction 0.89 0.31 **
Self-efficacy 0.63 0.37
Constant -0.70 0.36 -1.22 0.39** -3.63 0.61**
N
1,455 1,455 1,455
McFaddens R2 0.04 0.10 0.12
Notes: *; **: P < 0.05; 0.01 (two-tailed tests). The joint significance of categorical variables is tested using an F-test,
and the level of significance is indicated in the first row for each categorical variable. The table reports logit coefficients
and robust standard errors.
Model II includes education and control variables that are expected to be causally prior to
education. The results provide evidence in favor of H2, stating that higher education is associated
15
with higher social trust. In order to have an idea about the magnitude of the impact, we predicted
the average probability of trusting others at different levels of education, and this points to a strong
effect. The average predicted probability of expressing trust for a respondent with primary school as
highest education is 0.58, while it is 0.93 if s/he had an academic education.12 H3 is also partially
supported in Table 1 as the differences in trust between the cohorts are reduced when education is
introduced to the model, but they do not disappear entirely. This indicates that the cohort effect on
trust is in part due to generational differences in the level of education.
A strong association between institutional trust and social trust is evident from Model III. Despite
controlling for a large number of possible confounders, institutional trust has a significant impact on
social trust, which supports H4. The average predicted probability of trusting others increases from
0.56 to 0.87 when comparing the lowest to the highest level of institutional trust. Hence,
institutional trust has a marked impact on social trust, roughly corresponding to that of education
(based on Model II).
However, an association is not enough to substantiate a causal impact of institutional trust on social
trust as noted earlier, and to provide further support for the suggested interpretation of this
correlation, we examine the causal relationship between the two types of trust in Table 2.
Models IV and VI in Table 2 show that both forms of trust measured in 2008 are markedly
influenced by the same type of trust measured in 1999, and for institutional trust, trust in 2008 is
even influenced by trust in 1990. In relation to the present research question, the interesting finding
is that institutional trust in 1999 has a significant impact on social trust in 2008 despite controlling
for social trust measured in 1999 and 1990 and a large number of control variables. Moreover, we
do not find the reverse relationship, as there is no evidence for an effect of social trust measured in
1990 or 1999 on institutional trust in 2008. Similarly, in the models with only one time lag (Models
V and VII), we also find only an effect of institutional trust on social trust and not the reverse
relationship. This is evidence of institutional trust having a causal impact on social trust and not the
other way around. Furthermore, to underline the importance of institutional trust in forming social
trust, it should be noted that this is the only variable apart from lagged social trust and education
that significantly influences social trust (at the 0.05-level) in Models IV-V. Even associational
involvement, the previously prominent explanation of social trust, is insignificant and close to zero
once previous levels of social trust are controlled for. This finding is consistent with the self-
selection hypothesis, i.e. that social trusters are more likely to join voluntary associations (Stolle,
2001).
16
Table 2: Causes of social trust and institutional trust, panel data
Model IV Model V Model VI Model VII
Dependent variable Social trust t0 Social trust t0 Institutional trust t0 Institutional trust t0
Estimator Logit Logit OLS OLS
Coef. S.E. Coef. S.E. Coef. S.E. Coef. S.E.
Institutional trust t-1 2.90 1.27* 1.25 0.60* 0.41 0.06** 0.44 0.03**
Institutional trust t-2 -0.91 1.25 - 0.14 0.05** -
Social trust t-1 1.58 0.38** 1.52 0.19** -0.00 0.02 0.01 0.01
Social trust t-2 0.39 0.38 - 0.01 0.02 -
Education t0 *
Primary school Reference Reference Reference Reference
Vocational -0.96 0.50 -0.13 0.24 -0.00 0.02 0.00 0.01
Shorter post-
secondary -0.92 0.52 0.26 0.29 0.01 0.02 0.02 0.02
Medium-long post-
secondary 0.44 0.77 0.45 0.31 0.00 0.03 -0.01 0.02
Academic 0.58 1.16 1.70 0.64** 0.03 0.03 0.04 0.02*
Employed (yes) t0 0.99 0.53 0.40 0.27 0.01 0.02 -0.00 0.01
Relative household
income t0
< 25. Percentile Reference Reference Reference Reference
Interquartile range -0.40 0.54 0.06 0.29 0.00 0.02 -0.02 0.02
>75. Percentile 0.26 0.74 -0.10 0.29 -0.00 0.03 -0.00 0.01
Not revealed -1.17 0.50* -0.54 0.29 -0.01 0.02 0.01 0.02
Associational
involvement t0 0.39 0.36 0.30 0.19 -0.01 0.01 0.01 0.01
Life satisfaction t0 0.69 0.89 0.24 0.51 0.02 0.04 0.04 0.03
Self-efficacy t0 1.43 0.86 0.66 0.49 0.13 0.05** 0.09 0.03**
Constant -2.23 1.47 -1.51 0.61* 0.27 0.07** 0.29 0.03**
Time-invariant controls yes yes yes yes
N
324 786 322 786
McFaddens R2/R2 0.25 0.18 0.33 0.26
Notes: *; **: P < 0.05; 0.01 (two-tailed tests). The joint significance of categorical variables is tested using the F-test,
and the level of significance is indicated in the first row for each categorical variable. The table reports coefficients and
robust standard errors. t0, t-1, and t-2 refer to 2008, 1999, 1990, respectively, in Models IV and VI. In Model V and VII,
t0 and t-1 refer to either 2008 and 1999, or 1999 and 1990. Models IV-VII contain the same control variables as in Table
1, but the coefficients for the time-invariant variables birth cohort, gender, and place of birth are not shown.
17
Finally, a word on the insignificant effect of education on social trust in Model IV, which may be
seen as a weakening of H2. The insignificance is likely to be caused by little individual-level
variation in education over time. Since this is a sample of adults, the level of education does not
change much over time for most respondents, which implies that most of the effect of education is
probably captured by the lagged trust variable in Model IV.
In sum, the analyses generally support the four hypotheses put forward. Hence, our results indicate
that the increasing levels of trust over the past 30 years are the result of generational replacement
and concomitant rising levels of education as well as increasing levels of trust in institutions, which
appear to reflect the quality of state institutions.
Conclusion and discussion
Evidence suggests that the level of social trust in Denmark is (among) the highest in the world
today. Given that social trust is an important resource contributing to the provision of public goods
and promoting cooperation, this paper has sought to understand the background for this
exceptionalism and thus the factors that may work to increase trust more generally. We have
examined two generic types of explanations for the high level of trust: cultural and experiential
explanations. An analysis of the development in trust over the past 30 years showed that it Danish
trust exceptionalism is a recent phenomenon as trust has increased dramatically in Denmark over
the last 30 years. This clearly suggests that the deep historic factors embodied in the cultural
perspective on trust cannot explain the elevated present-day level of trust in the Danish case. Instead
our results suggest that the increasing levels of trust, and hence the current trust exceptionalism, can
best be understood in experiential terms as a combination of cohort and period effects. More
specifically, we find that generational replacement explains part of the increase in social trust as
trust is significantly lower among older cohorts. Moreover, our analyses show that the differences in
trust between cohorts can, to a considerable extent, be explained by differences in educational
levels. Finally, we also found that (increased) institutional quality is the most powerful period effect
explaining the increase in trust.
These results are good news for benevolent policy-makers and others who strive to grow social
trust, as our results clearly indicate that social trust is indeed manipulable if not in the short term
then at least in the medium term. If levels of social trust were mainly the product of forces set in
motion centuries ago, this would leave reformers without tools to promote trust. Given the positive
influence of education and institutional quality mediated through institutional trust our results
18
suggest that promoting higher education and building up well-functioning state institutions can
foster social trust. Importantly, the present analysis also provides rather strong evidence that
causality actually runs from institutional trust to social trust and not the other way around. This
suggests that fair and efficient state institutions promote social trust, and that improving the quality
of state institutions probably is a viable strategy in order to uphold and potentially build up the
valuable resource of social trust. This is a rather different strategy than a previously prominent
receipt for promoting trust, namely that of furthering associational engagement, which, our results
show only have a limited effects on social trust. In that sense, our results add to the growing pile of
evidence suggesting that institutions rather than civic life are the main driver of social trust.
The increase in social trust among Danes is striking in a comparative perspective, but the results of
the analysis of the causes of this development are also interesting when compared to within-country
analyses from other countries. Most remarkably, the results reveal reverse consequences of
generational replacement for social trust in Denmark and the US. The present analysis showed that
the oldest cohorts are less trustful than the younger cohorts in Denmark. The opposite is found in
the US, where cohorts born approximately before the end of World War II are more trustful than
younger cohorts, which explains a large part of the much discussed drop in social trust in this
country (Putnam, 1995; Robinson & Jackson, 2001). Consequently, the level of trust among the
cohorts born before 1945 appears to have been largely similar in the two countries, perhaps with a
small lead to the US (Putnam, 1995: 675). This suggests that the divergent trends in trust between
the two countries were founded among the cohorts born after the war. This naturally leads to the
question about which differences in the post-war cohorts’ environments in the two countries that
may account for these divergent trends. Given that the two countries obviously differ in a great
number of ways, one can only speculate about this explanation. On the face of it, education would
not be able to explain this divergence as education has generally been increasing in the second part
of the 20th century in both countries. However, it is worth noting that in contrast to the development
in Denmark, some indicators suggest that the increase in the level of education in the US has
attened out since the mid-1970s (U.S. Census Bureau, 2012), as also noted by Robinson &
Jackson (2001). In other words, differences in the development in education may be a plausible
explanation for the divergent trends in trust in the two countries after all. Another plausible
explanation is that the cohorts growing up during the Vietnam War and the civil rights movement in
the USA had collective experiences early in life affecting trust negatively (Uslaner, 2002), which
are unparalleled in Denmark.
19
Notes
1 According to this definition, social trust is distinct from other forms of trust in that it does not concern perceptions of
specific others or specific groups. The latter types of trust are often referred to as particularized trust and group- or
identity-based trust, respectively. See the discussions in Freitag & Bauer (2013), Patulny & Svendsen (2007),
Sønderskov (2008) and Uslaner (2002).
2 However, with Iceland scoring markedly lower than the other countries.
3 As an alternative to the cultural explanation explaining intergenerational stability in trust by parental socialization
early in life, it should be noted that recent research has shown that the stability may also be the result of genetic
transmission of trust from parents to children (Sturgis et al., 2009; Hirashi et al., 2008). However, for the present
purposes, the important issue is that both perspectives predict stability in trust.
4 Note that parts of this figure (and Figure 2 and 3) have also been reported in Dinesen & Sønderskov (2012b).
5 Note that the wording of the trust question in the American General Social Survey is marginally different than in the
remaining surveys. The wording is "Generally speaking, would you say that most people can be trust or that you cannot
be careful in life".
6 The literature on value change typically also consider age explanations, which emphasize how values develop with
age over the life course according to changes in life circumstances and position in society. Generally speaking, all
individuals are expected to experience the same value change with age (e.g., the transition from being left-wing during
youth to becoming more right-wing with age). Unless a society’s age composition changes markedly over time, age
explanations are generally not suited for explaining value change at the societal level, as the expectation from this
perspective is that values are stable at the aggregate level over time as everyone is expected to go through the same
value change over the life course. Similar to most developed countries, the age composition has been relatively stable
over the last 30 years in Denmark (Statistics Denmark, 2008), and thus, we can disregard age effects as an explanation
for the increase in trust.
7 Statistics Denmark, account BEF1A.
8 Note, however, that the reliability and validity of the institutional quality measure have been criticized (Kurtz &
Schrank, 2007) and inter-temporal comparison may therefore be somewhat problematic.
9 The full 1990-2008 panel dataset (Danish Data Archive 2009) only includes respondents who chose to participate in
all three waves. We have merged this dataset with the 1990-1999 dataset (Danish Data Archive 2002) to retain
respondents who did not participate in 2008.
10 The robustness of the impact of institutional trust on social trust is assessed in a number of ways: omitting
associational involvement from the model, controlling for an alternative measure of involvement, dropping respondents
affected by non-response in the household income variable, omitting household income from the model (see notes to
Supplementary Table A). The coefficient for institutional trust does not change in these models (not shown).
11 To avoid losing too many observations in the analysis because of non-response on one of the questions about trust in
institutions, we included respondents who responded to at least three of the questions. The index has a Cronbach’s alpha
between 0.69 and 0.74 in the cross-sectional data and the three waves of the panel.
12 While our results point to a strong effect of education on social trust, it should be noted that recent research have
questioned whether this is a causal relationship or rather that it reflects (is confounded by) various unobserved factors
(Oskarsson et al., 2014).
20
References
Barro, Robert & Lee, Jong-Wha. (2010). A New Data Set of Educational Attainment in the World,
1950-2010. NBER Working Paper No. 15902. Downloaded December 21, 2011 from
http://www.barrolee.com/data/BL_v1.2/BL(2010)_MF2599_v1.2.dta.
Bekkers, René. (2012). Trust and Volunteering: Selection or Causation? Evidence from a 4 Year
Panel Study. Political Behavior, 34(2), 225-247.
Bjørnskov, Christian. (2006). Determinants of generalized trust: A cross-country comparison.
Public Choice, 130(1), 1-21.
Bjørnskov, Christian. (2008). Social Trust and Fractionalization: A Possible Reinterpretation.
European Sociological Review, 24(3), 271-283.
Brehm, John & Rahn, Wendy. (1997). Individual-Level Evidence for the Causes and Consequences
of Social Capital. American Journal of Political Science, 41(3), 999-1023.
Clark, April K. & Eisenstein, Marie A. (2013). Interpersonal Trust: An Age-Period-Cohort Analysis
Revisted. Social Science Research, 42(2), 361-375.
Coleman, James S. (1990). Foundations of Social Theory. Cambridge: Harvard University Press.
Danish Data Archive. (1981). Study DDA-228: Politiske værdier i Danmark 1979. Obtained from
the Danish Data Archive.
Danish Data Archive. (2002). Study DDA-9991: International værdiundersøgelse (Danmark): Panel
1990-1999”. Obtained from the Danish Data Archive.
Danish Data Archive. (2009). Den danske værdiundersøgelse, panel 1990-2008”. Obtained from the
Danish Data Archive.
Danish Data Archive. (2011). Study DDA-24772: Social ulighed i Danmark (ISSP 2009). Obtained
from the Danish Data Archive.
Delhey, Jan & Newton, Kenneth. (2003). Who trusts?: The origins of social trust in seven societies.
European Societies, 5(2), 93-137.
Delhey, Jan & Newton, Kenneth. (2005). Predicting Cross-National Levels of Social Trust: Global
Pattern or Nordic Exceptionalism?, European Sociological Review, 21(4), 311-327.
Dinesen, Peter Thisted. (2012a). Parental Transmission of Trust or Perceptions of Institutional
Fairness? Generalized Trust of Non-Western Immigrants in a High Trust Society. Comparative
Politics, 44(3), 273-289.
Dinesen, Peter Thisted (2012b). Does Generalized (Dis)Trust Travel? Examining the Impact of
Cultural Heritage and Destination-Country Environment on Trust of Immigrants. Political
Psychology, 33(4), 495-511.
Dinesen, Peter Thisted. (2013). Where You Come From or Where You Live? Examining the
Cultural and Institutional Explanation of Generalized Trust Using Migration as a Natural
Experiment. European Sociological Review, 29(1), 114-128.
Dinesen, Peter Thisted & Sønderskov, Kim Mannemar. (2012a). Trust in a time of increasing
diversity: On the relationship between ethnic heterogeneity and social trust in Denmark from 1979
until today. Scandinavian Political Studies, 35(4), 273-294.
21
Dinesen, Peter Thisted & Sønderskov, Kim Mannemar. (2012b). Hvorfor stiger tilliden?. Politica,
44(1), 87-110.
Dinesen, Peter Thisted, & Sønderskov, Kim Mannemar. (2013). Ethnic Diversity and Social Trust:
The Role of Exposure in the Micro-Context. Paper presented at Ethnic Diversity and Social Capital,
Berlin, May 24-25, 2013.
EVS. (2011). Longitudinal file 1981-2008; Pre-release v.1.0.0 as of April 30, 2011. Downloaded
December 9, 2011 from http://dx.doi.org/10.4232/1.4804.
Fairbrother, Malcolm, & Martin, Isaac W. (2013). Does inequality erode social trust? Results from
multilevel models of US states and counties. Social Science Research, 42(2), 347-360.
Finansministeriet. (2010). Indkomstudvikling og -fordeling i Danmark 1983-2007. Downloaded
December 21, 2011 from http://www.fm.dk/Nyheder/Pressemeddelelser/2010/03/~/media/Files
/Nyheder/Pressemeddelelser/2010/03/Indkomstudvikling%20og%20fordeling%20i%20danmark%2
083%20til%2007/indkomstudvikling_og_fordeling_i_danmark_83_07.ashx.
Freitag, Markus, & Bauer, Paul C. (2013). Testing for measurement equivalence in surveys
dimensions of social trust across cultural contexts. Public Opinion Quarterly, 77(S1), 24-44.
Gächter, Simon, Benedikt, Herrman & Thöni, Christian. (2004). Trust, voluntary cooperation, and
socio-economic background: survey and experimental evidence. Journal of Economic Behavior and
Organization, 55(4), 505-531.
Glanville, Jennifer L., & Paxton, Pamela. (2007). How do we learn to trust? A confirmatory tetrad
analysis of the sources of generalized trust. Social Psychology Quarterly, 70(3), 230-242.
Glenn, Norval D. (2005). Cohort analysis, Second eds. Thousand Oaks: Sage.
GSS (2011) “GSS 2010 merged with all cases and variables (Release 1, Oct. 2011)”, Downloaded
December 9, 2011 from http://www3.norc.org/GSS+Website/Download/STATA+v8.0+Format/.
Hirashi, Kai, Yamagata, Shinji, Shikishima, Chizuru & Ando, Juku. (2008). Maintenance of genetic
variation in personality through control of mental mechanisms: a test of trust, extraversion and
agreeableness. Evolution and Human Behavior, 29, 79-85.
Kaufmann, Daniel, Kraay, Aart & Mastruzzi, Massimo. (2011). The Worldwide Governance
Indicators (WGI) project. The World Bank. Downloaded December 21, 2011 from
http://info.worldbank.org/governance/wgi/index.asp.
Knack, Stephen & Keefer, Phillip. (1997). Does social capital have an economic payoff? A cross-
country investigation. Quarterly Journal of Economics, 112(4) 1251-1288.
Knack, Stephen. (2002). Social capital and the quality of government: Evidence from the states.
American Journal of Political Science, 46(4), 772-785.
Kurtz, Marcus J., & Schrank, Andrew. (2007). Growth and governance: Models, measures, and
mechanisms. Journal of Politics, 69(2), 538-554.
Nannestad, Peter, Svendsen, Gert Tinggard, Dinesen, Peter Thisted, & Sønderskov, Kim
Mannemar. (2014). Do Institutions or Culture Determine the Level of Social Trust? The Natural
Experiment of Migration from Non-western to Western Countries. Journal of Ethnic and Migration
Studies, 40(4), 544-565.
22
Oskarsson, Sven, Dawes, Chris T., Dinesen, Peter Thisted, Johannesson, Magnus, & Magnusson,
Patrik K. E. (2014). Education and Social Trust: Testing a Causal Hypothesis Using the Discordant
Twin-Pair Design. Mimeo.
Patulny, Roger V. & Svendsen, Gunnar Lind Haase (2007) Exploring the social capital grid:
bonding, bridging, qualitative, quantitative, International Journal of Sociology and Social Policy,
27(1/2), 32-51.
Putnam, Robert D. (1993). Making Democracy Work: Civic traditions in modern Italy. Princeston:
Princeton University Press.
Putnam, Robert D. (1995). Tuning In, Tuning Out: The Strange Disappearance of Social Capital in
America. PS. Political Science and Politics, 28(4), 664-683.
Putnam, Robert D. (2000). Bowling alone: The Collapse and Revival of American Community. New
York: Simon & Schuster.
Rice, Tom W. & Feldman, Jan L. (1997). Civic culture and democracy from Europe to America.
Journal of Politics, 59(4), 1143-1172.
Robbins, Blaine G. (2012). Institutional Quality and Generalized Trust: A Nonrecursive Causal
Model. Social Indicators Research, 107(2), 235-258.
Robinson, Robert V. & Jackson, Elton F. (2001). Is Trust in Others Declining in America? An Age–
Period–Cohort Analysis. Social Science Research, 30, 117-145.
Rostila, Mikael. (2007). Social capital and health in European welfare regimes: a multilevel
approach. Journal of European Social Policy, 17, 223-239.
Rothstein, Bo & Stolle, Dietlind. (2008). The State and Social Capital. An Institutional Theory of
Generalized trust. Comparative Politics, 40(4), 441-60.
Rothstein, Bo & Uslaner, Eric. (2005). All for All: Equality, Corruption, and Social Trust. World
Politics, 58(1) 41-72.
Rönnerstrand, Björn. (2013). Social capital and immunisation against the 2009 A (H1N1) pandemic
in Sweden. Scandinavian Journal of Public Health, 41(8), 853-859.
Scholz, John T. & Lubell, Mark (1998). Trust and taxpaying: Testing the heuristic approach to
collective action. American Journal of Political Science, 42(2), 398-417.
Schwadel, Philip & Stout, Michael (2012). Age, Period, and Cohort Effects on Social Capital.
Social Forces, 91, 233-252.
Sears, David O. & Levy, Sheri. (2003). Childhood and Adult Political Development, pp. 60-109 in
Sears, David, Huddy, Leonie, Jervis, Robert (eds.). Oxford Handbook of Political Psychology. New
York: Oxford University Press.
Statistics Denmark (2008). 60 år i tal – Danmark siden 2. verdenskrig. Copenhagen: Statistics
Denmark.
Stolle, Dietlind. (2001). Clubs and Congregations: The Benefits of Joining an Association, pp. 202-
244 in Karen Cook (ed.). Trust in society. New York: Russell Sage Foundation.
Sturgis, Patric, Read, Sanna, Hatemi, Peter K., Zhu, Gu, Trull, Tim, Wright, Margaret J. & Martin,
Nicholas. (2009). A Genetic Basis for Social Trust?. Political Behavior, 32(2), 205-230.
23
Sønderskov, Kim Mannemar (2008). Making cooperation work: Generalized social trust and large-
N collective action. Aarhus: Politica.
Sønderskov, Kim Mannemar. (2011a). Explaining Large-N Cooperation: Generalized Social Trust
and the Social Exchange Heuristic. Rationality and Society, 23(1), 51-74.
Sønderskov, Kim Mannemar. (2011b). Does Generalized Social Trust Lead to Associational
Membership? Unravelling a Bowl of Well-Tossed Spaghetti. European Sociological Review, 27(4),
419-434.
Tabellini, Guido (2008). Presidential Address. Institutions and Culture. Journal of the European
Economic Association, 6(2-3), 255-294.
US Census Bureau (2012). Educational Attainment in the United States: 2009. U.S. Census Bureau.
Uslaner, Eric. (2002). The Moral Foundation of Trust. New York: Cambridge University Press.
Uslaner, Eric. (2008). Where You Stand Depends On Where Your Grandparents Sat: The
Inheritability of Generalized Trust. Public Opinion Quarterly, 72(4), 725-740.
Wang, Lanlan & Gordon, Peter. (2011). Trust and institutions: A multilevel analysis. Journal of
Socio-Economics, 40(5), 583-593.
WVS (2009). WVS official 5 wave aggregate 1981-2008 v.20090914. Downloaded September 15,
2011 from http://www.wvsevsdb.com/wvs/WVSData.jsp.
24
SupplementaryTableA:Codinganddescriptivesforallvariables
VariableCodingMean/Std.devordistribution
 CrosssectiondataPanelwave1Panelwave2Panelwave3
Socialtrust
“Generallyspeaking,doyouthinkthatmostpeoplecanbe
trustedorthatyoucannotbetoocarefulindealingwith
otherpeople?”withtworesponsecategories:“Mostpeople
canbetrusted”or“Youcannotbetoocareful”
“Mostpeoplecanbe
trusted”:76%
“Mostpeoplecanbe
trusted”:71%
“Mostpeoplecanbe
trusted”:74%
“Mostpeoplecanbe
trusted”:81%
Birthcohort
Decadeofbirth(pre19301980’s);
respondentsbornbefore1930arecollapsedinasingle
category
Pre1930:4%,30’s:10%,
40’s:19%,50’s:19%,
60’s:21%,70’s:15%,
80’s:12%
Pre1930:5%,30’s:13%,40’s:22%,50’s:28%,
60’s:28%,70’s:5%,80’s:0%
Education
Fivecategories:1)primaryschool,2)vocationaleducation,
3)shorterpostsecondaryeducation(12yearsposthigh
school),4)mediumlongpostsecondaryeducation(34
yearsposthighschool),and5)academiceducation(more
than4yearsposthighschool)
1:22%,2:30%,
3:21%,4:17%,5:10%
1:23%,2:35%,3:
23%,4:13%,5:7%
1:21%,2:36%,3:
20%,4:14%,5:9%
1:20%,2:31%,3:
26%,4:15%,5:9%
Institutional
trust
Indexwithrange01basedonitemsontrustinthe
parliament,thejudiciary,thepolice,andthepublicsector0.63/0.140.60/0.160.59/0.150.65/0.15
Gender(female)Dummyvariablewhere1indicatesfemaleFemale:50%Female:49%
Placeofbirth
(native)
Dummyvariablewhere1indicatesnative(asopposedto
foreignborn)Native:94%Native:97%
Household
income
Threecategoriesindicatingtherelativelevelofhousehold
incomecomparedtotheotherrespondentsinthesurvey
(belowthe25.percentile,withintheinterquartilerange,or
abovethe75.percentile)andthefourthcategorybeingnon
responsea
1:19%,2:41%,
3:14%,4:26%
1:17%,2:46%,3:
25%,4:11%
1:24%,2:44%,3:
27%,4:5%
1:21%,2:39%,3:
22%,4:19%
Associational
involvement
Dummyvariablewhere1indicatesthattherespondentdid
voluntaryworkinatleastoneof15
organizations/associationsatthetimeofthesurveyb
Involved:34%Involved:37%Involved:52%Involved:43%
Lifesatisfaction“Howsatisfiedordissatisfiedareyouwithlifenowadays?”0.82/0.200.83/0.180.83/0.150.84/0.18
25
10responsecategoriesfromzero(“Verydissatisfied”)to1
(“Verysatisfied”)
Selfefficacy
“Howmuchinfluencedoyouhaveoveryourownlife?”10
responsecategoriesfromzero(”Noinfluenceatall”)to1(”A
lotofinfluence”)
0.75/0.190.70/0.180.71/0.190.73/0.18
Notes:Thedescriptivesforthecrosssectionaldataisbasedonthe1,455respondentsincludedinModelsIIII.Thedescriptivesforthepaneldataisbasedonthe324respondentsincludedinModel
IV.a:Nonresponse(includingthosenotwishingtoanswerthequestion)wasincludedasacategorytoavoidexcludinganumberofrespondentsfromtheanalysis.Asnonresponsemaybeaffected
bytrust(andhencetheriskofsimultaneitybias),wealsoestimatedmodelsomittingthenonresponsecategories.b:Weoptedforthisoperationalizationasvoluntaryworkmorelikelyinvolves
interactionwithotherpeoplethanmeremembership(whichmaybepassive)andhencecomesclosertothemainmechanismlinkingparticipationinorganizationstotrust(Sønderskov,2011b).As
analternativemeasure,wetriedusingasummatedscaleofthenumberofvoluntaryassociationsinwhichtherespondentcarriedoutvoluntarywork.
... While the literature exhibits uncertainties regarding the causal direction between these various forms of trust (Newton and Zmerli 2011) it generally suggests that trust in institutions spills over to trust in people in general. In an extensive study, Sønderskov and Dinesen (2016), using data from the hightrust country of Denmark, found that trust in state institutions has a causal impact on generalised social trust, while the evidence for a reverse relationship is limited (see also Sønderskov and Dinesen 2014). Similar findings are reported by Seifert (2018) in the Netherlands and Switzerland. ...
... This argument is supported by a number of cross-sectional studies that focus on the relationship between generalised social trust and political trust (Dinesen 2013;Freitag and Bühlmann 2009;Tao et al. 2014). In one of the most extensive studies based on two Danish panel studies, Sønderskov and Dinesen (2016) found that trust in state institutions had a causal impact on generalised social trust, while the evidence for an inverse relationship was limited (see also Sønderskov and Dinesen 2014). A similar conclusion is presented by Seifert (2018) using data from the Netherlands and Switzerland. ...
... Our findings differ from previous research based on the experiential perspective, which suggests that trust in institutions naturally extends to trust in people in general (e.g., Seifert 2018;Sønderskov and Dinesen 2014). Instead, our study contributes to the prior research by showing that in a local context, where daily interactions take place between community members and representatives of local institutions, the relationship between trust in neighbours and trust in local institutions appears to be more complex and interdependent. ...
Article
Full-text available
Decades of social science stress the vital role of trust in a robust democracy, and scholars have explored the origins of trust, highlighting the dynamic interaction between institutional trust and generalised social trust. While the literature exhibits uncertainties regarding the causal direction between these various forms of trust, it generally suggests that trust in institutions spills over to trust in people in general. This article examines a neglected aspect of trust studies-the relationship between local institutional trust and community trust. Drawing on a two-wave panel study from Sweden, the results indicate bidirectional relationships between the two forms of trust over time. In so doing, this study contributes to the prior research by showing that in a local context, where daily interactions take place between community members and representatives of local institutions, the relationship between trust in neighbours and trust in local institutions appears to be more complex and interdependent.
... ‫کردار‬ ‫و‬ ‫گفتار‬ ‫به‬ ‫اعتماد‬ ‫و‬ ‫خیر‬ ‫نیت‬ ‫دادن‬ ‫نسبت‬ ‫به‬ ‫تمایل‬ ‫میزان‬ (COOK & WALL, 1980) ‫دیگری.‬ ‫به‬ ‫فرد‬ ‫وابستگی‬ ‫آگاهانه‬ ‫تنظیم‬ (Zand, 1972) (Sonderskov & Dinesen, 2014) . (Locke, 1976) . ...
Article
Full-text available
Trust has penetrated almost in all social aspects of human life and especially in the heart of organizational interactions, and as a glue that holds society together, it has attracted significant interest from management thinkers. Trust, both within and between organizations, is an important facilitator of cooperation and can increase organizational effectiveness. Trust is very dynamic and is influenced by various factors, especially cognitive factors, which unfortunately have steadily decreased in recent years. Familiarity with cognitive factors affecting the way trust works is a necessity that is investigated in this research. This study has been conducted with the combined method of inductive-deductive approach and specifically through meta-combination of 32 articles to identify cognitive factors effective in organizational trust. Then these factors were prioritized through a survey of 12 experts and specialists who were selected with a targeted sampling approach. Expert survey data were used to analyze quantitative data and understand the cognitive factors affecting organizational trust, using the best-worst method (BWM). The results based on the examination of theoretical foundations provided the insight that organizational trust is influenced by 10 cognitive factors. Then by prioritizing the mentioned factors, "honesty" was identified as the most important (best) factor and "knowledge sharing" as the least important (worst) factor. This article presented a cognitive approach to the antecedents of organizational trust and highlighted the role of cognitive antecedents in relation to organizational trust.
... Rothstein and Stolle (2008), for example, find a positive relationship between trust in institutions such as police and civil service and social trust. Sønderskov and Dinesen (2014), Sønderskov and Dinesen (2016), and Seifert (2018) strengthen this claim by using panel data and showing that institutional trust strongly and positively influences social trust. ...
... The biggest issue with introducing workplace monitoring among managers and employees was the concern about the decrease in trust, which is especially an issue in Denmark as a country of high trust (Sønderskov & Dinesen, 2014). If workplace monitoring should be introduced, it is necessary to balance trust and surveillance and thus include all stakeholders from the beginning -employees, managers, HR, and union representatives. ...
Article
Full-text available
Many knowledge workers have shifted to work hybrid by being partly in the office and partly at remote locations. This brings freedom for knowledge workers, but managers are concerned about the performance of their workforce when they are not working from the office. One means to check performance constantly is by collecting and analyzing performance and behavior data to monitor employees, which is a form of algorithmic control. On the one hand, this can lead to conflicts between employees and managers as employees feel that managers mistrust them. On the other hand, employees might trust transparent monitoring that helps them focus on core tasks. To learn more about balancing trust and surveillance, this pilot study surveyed 45 knowledge workers in an engineering company in Denmark, a country with high levels of trust in society. We developed an online survey to investigate managers’ and employees’ opinions about workplace monitoring. We found that all participants had a negative to neutral attitude toward monitoring, although managers were slightly more favorable. Furthermore, data collection for work-related purposes to improve the work situation might be appropriate, although a trustful relationship between managers and employees is the most important. Thus, if a form of monitoring is going to be introduced at the workplace, it is essential that all stakeholders – managers, employees, HR, and union representatives – are involved and that complete transparency is achieved in terms of which data is collected and for which purposes. Only then can a balance between trust and surveillance be achieved, including maintaining a positive working climate.
Article
Public services often require formal documentation from citizens or public employees. Although these administrative requirements are commonly viewed as burdensome, they play a critical role in safeguarding program integrity. Drawing from a pre‐registered survey experiment conducted among Danish residents ( n = 2004), this article examines citizen attitudes toward policy reforms that either intensify or alleviate the burden of formal documentation requirements. We analyze how these contrasting types of policy reforms shape citizen support, with a particular emphasis on the trade‐off between compliance burden and program integrity protection. Furthermore, the impact of whether these changes in requirements apply to citizens or public employees is explored. Our findings suggest a general preference among citizens for the removal of administrative documentation requirements aimed at protecting program integrity, regardless of whether they apply to citizens or public employees. These attitudes appear influenced by political ideology, personal experience, and the specific policy area context.
Article
This paper explores the long‐term impact of China's Great Leap Forward (GLF) on current differences in cooperative behaviour across Chinese regions and the underlying trust channel of causality. Combining the current firm‐level survey with grain yield over‐reporting during the GLF era, we find that past government dishonesty regarding over‐reporting in the GLF period has long‐term negative consequences on people's behaviours. The firms in the prefecture with a higher degree of yield over‐reporting are less likely to engage in R&D collaborations today. Further tests reveal the possible trust channel through which past government dishonesty impedes current cooperative behaviour. We extend the analyses to social norms of trust and show that GLF over‐reporting has shifted an individual's internal norms towards a present‐day mistrust culture, which contributes to explain the negative effect of GLF over‐reporting on cooperative behaviour. Our findings suggest that dishonest behaviours by local governments in the past can lead to lasting adverse effects on an individual's behaviour in the form of mistrust and subsequent lower frequencies of cooperative behaviour, providing new insights on the origins of cultural differences and human behaviours by investigating the role of governments in shaping social and economic interactions.
Article
Full-text available
Danmark er berømt for høj mellemmenneskelige tillid – tillid til andre mennesker generelt – eller hvad der i forskningslitteraturen benævnes (generaliseret) social tillid. Den høje sociale tillid betragtes som en væsentlig samfundsressource, fordi den fremmer samarbejde mellem folk, der ikke kender hinanden. På samfundsplan manifesterer den høje tillid sig bl.a. ved, at flere er tilbøjelige til at udvise samfundssind og på andre måder opføre sig uegennyttigt, fordi de forventer – stoler på – at andre omkring dem gør det samme. Høj social tillid er således sandsynligvis en central forklaring på, hvorfor Danmark er et relativt velfungerende samfund. På den baggrund er det relevant at spørge, om tilliden er truet? I dette essay gør vi status over tilliden i Danmark og diskuterer potentielle trusler imod den.
Article
Full-text available
Social trust is a deeply-rooted feature of society, whose positive impact on economic performance has been widely documented in many contexts. However, its impact on the non-economic aspects of social progress that characterize advanced societies, such as personal rights, freedom, tolerance and inclusion and access to advanced education is still understudied, especially at the subnational level. As shown by the European Social Progress Index (EU-SPI) 2020, elaborated by the European Commission, the European regions present remarkable disparities in those non-economic aspects. Using the EU-SPI framework, this paper provides fresh evidence on a positive impact of social trust on several features defining advanced social progress. Social trust effects are mainly seen in improved quality of government, education and people's pro-social behaviors. These insights can be useful for the design of future policies that pursue a more equal Europe beyond purely economic indicators, given that regional social trust can condition their success.
Article
Full-text available
Purpose The purpose of this paper is to show that numerous studies have advanced social capital research over the past decade. Most studies have accepted the theoretical distinction between bonding and bridging social capital networks. Many, however, tend to agglomerate empirical research under the one catch‐all social capital concept, rather than classifying it according to the bonding/bridging distinction. Furthermore, most studies make little distinction on the basis of methodology, between qualitative and quantitative approaches to investigating social capital. These omissions need to be addressed. Design/methodology/approach This paper reviews definitions and applications of bridging and bonding social capital, classifies empirical studies according to each network type, and produces a further breakdown according to methodological approach. Findings The result is a four‐part “grid” of social capital research, encompassing bonding and bridging, and quantitative and qualitative aspects. This paper finds that most qualitative research examines non‐excludable and excludable goods and is relevant to bonding social capital, whilst most quantitative analysis looks at civic networks and norms of trust, and relates to bridging social capital. Research limitations/implications Results advance the task of clarifying and measuring social capital. Practical implications Further development of the bridging/bonding social capital conceptual pair should allow for a more precise measurement of a community, or region. Originality/value No review paper to date captures the above empirical and methodological “grid” clearly.
Article
Full-text available
In the discussion of the sources of social capital, it has been stressed that generalized trust is built up by the citizens themselves through a culture that permeates the networks and organizations of civil society. This approach has run into conceptual problems, and empirical evidence has provided only mixed support. An alternate approach is to highlight how social capital is embedded in and linked to formal political and legal institutions. Not all political institutions matter equally, however. Trust thrives most in societies with effective, impartial, and fair street-level bureaucracies. The causal mechanism between these institutional characteristics and generalized trust is illustrated in a cross-national context.
Article
One of the clearest results in previous studies on social trust is the robust positive relationship with educational attainment. The most common interpretation is that education has a causal effect on social trust. The theoretical argument and empirical results in this article suggest a different interpretation. We argue that common preadult factors such as cognitive abilities and personality traits rooted in genes and early-life family environment may confound the relationship between educational attainment and social trust. We provide new evidence on this question by utilizing the quasi-experiment of twinning. By looking at the relationship between education and social trust within monozygotic (MZ) twin pairs, we are able to avoid potential confounders rooted in genetic factors and common environmental influences because the monozygotic twins share both. The results suggest that when controlling for such familial factors the estimated effects of education on social trust are close to zero and far from reaching statistical significance. Further analyses show that the relationship between education and social trust largely is driven by common genetic factors.
Article
By utilizing the natural experiment of migration, this article attempts to answer whether generalized trust in other people is the result of cultural heritage or institutional quality. Looking at immigrants having migrated from a broad range of countries of origin to destination countries in Western Europe, I examine how their generalized trust is affected by the culture of their country of origin (in terms of the level of trust of this country) as well as institutional quality in the country they have migrated to (in terms of freedom from corruption). The results show that controlling for confounding variables, both factors have a highly significant impact on trust and hence that generalized trust appears to have both cultural and institutional foundations. © 2011 The Author. Published by Oxford University Press. All rights reserved.
Article
Social capital is the web of cooperative relationships between citizens that facilitates resolution of collection action problems (Coleman 1990; Putnam 1993). Although normally conceived as a property of communities, the reciprocal relationship between community involvement and trust in others is a demonstration of social capital in individual behavior and attitudes. Variation in social capital can be explained by citizens' psychological involvement with their communities, cognitive abilities, economic resources, and general life satisfaction. This variation affects citizens' confidence in national institutions, beyond specific controls for measures of actual performance. We analyze the pooled General Social Surveys from 1972 to 1994 in a latent variables framework incorporating aggregate contextual data. Civic engagement and interpersonal trust are in a tight reciprocal relationship, where the connection is stronger from participation to interpersonal trust, rather than the reverse.
Article
Theory: Trust is a critical attitude that extends the duty heuristic developed in Scholz and Pinney (1995). The "trust heuristic" can provide the basis for a contingent compliance strategy capable of sustaining cooperative solutions to collective action problems of governance if two conditions are met. First, compliance with laws must be conditional on levels of trust in specific legal arenas. Second, a citizen's trust in government and trust in other citizens' willingness to obey the law must reflect the costs and benefits associated with obeying laws. Hypothesis: This article tests the first hypothesis in the tax arena: trust in government and in other citizens increase compliance over and above the levels expected from an internalized sense of duty to obey laws and the fear of getting caught by enforcement agencies like the IRS. Method: We test the hypotheses with regression analysis of survey and tax return data from a stratified sample of 299 middle- and upper-income taxpayers, using the newly-developed two-stage conditional maximum likelihood analysis to control for endogeneity. We extend this approach to the analysis of multi-categorical ordered dependent variables. Findings: Both dimensions of trust significantly increase the likelihood of tax compliance, even after controlling for duty, fear, selection bias, and potential endogeneity effects.