Special Issue Paper
The impact of generalized and institutional trust on donating to activist, leisure, and interest organizations: individual and contextual effects
Maurice Gesthuizen, Radboud University Nijmegen, Department of Sociology, PO BOX 9104, 6500 HE, Nijmegen, The Netherlands.
In this paper, we answer the question as to what extent donating to activist, interest, and leisure organizations is affected by both individual and national levels of generalized and institutional trust. We use the European Social Survey 2002 to estimate multilevel random intercept models, based on more than 33 000 individuals living in 19 European countries and the USA. Our results show very consistent positive impacts of both individual generalized trust and institutional trust on donating to all types of organizations. The effects are strongest for donating to activist organizations and absent only for the relation between institutional trust and donating to interest organizations. At the national level, generalized trust positively affects donations to activist and leisure organizations, but not to interest organizations. Institutional trust at the national level is negatively related to donations to all types of organizations. This latter finding suggests that when institutions are perceived to function well, individuals estimate that their philanthropic donations are less needed. Copyright © 2011 John Wiley & Sons, Ltd.
A large branch of literature has emerged, studying the causes and consequences of social capital. An important reason is that in sociological, political, and economic research, social capital is acknowledged to be a key factor in generating better societies. High levels of trust among citizens and a population that is highly active within formal and informal circles are believed to, among others, produce lower crime rates, better democracies, and higher economic growth (Putnam, 2000; Halpern, 2005; yet see Portes and Vickstrom, 2011 for a critical review on this line of research).
Scholars use definitions of social capital as either the resources present within social networks of individuals and families that are accessible and transposable into individual gain (Bourdieu, 1983; Coleman, 1988) or as the amount of participatory potential, civic orientation, and trust in others, developed within and available to larger communities (Putnam, 2000). As the number of studies on the topic has so vastly expanded, the operational definitions of (dimensions of) social capital did so as well. A recent comparative research line focuses at explaining cross-national differences in formal and informal social capital (e.g., Gesthuizen et al., 2008; Pichler and Wallace, 2007; Savelkoul et al., 2011; Gesthuizen et al., 2009). Formal social capital pertains to activities employed within or on behalf of formally constituted civic organizations, such as donating to, being a member of, or donating to voluntary associations (Putnam, 2000; Schofer and Fourcade-Gourninchas, 2001), whereas informal social capital pertains to social ties between individuals and their friends, families, colleagues, and neighbors (Bourdieu, 1983; Burt, 2001; Coleman, 1988; Lin, 1999). In this line of research, giving to charitable organizations is explicitly considered as part of the formal dimension of social capital, as it pertains to the “civic orientation” of larger communities (Putnam, 2000) such as nations. Studies in philanthropy consider donating to charitable organizations as a form of prosocial behavior—and not as part of the concept of social capital—that can be explained by an individual's social capital, indicated for instance by associational networks of individuals and trust in others and their communities (Brown and Ferris, 2007), thereby relating more to the definition of social capital by Bourdieu (1983) and Coleman (1988).
As the purpose of our paper is to study the relationship between individual and national level trust and donating behavior in 19 European countries and the USA, our findings can be related to both literatures. Explaining why in some countries more citizens donate than in others, typically relates to the comparative literature described previously that addresses the civic orientation of nations. Yet, explaining why some individuals are more likely to donate than others, on the basis of individual differences in trust, can be placed in the philanthropic research branch. In this study, we distinguish between donating to activist, interest, and leisure organizations and between generalized and institutional trust. Our main research question is to what extent are individual and national levels of generalized and institutional trust related to donating of individuals to activist, leisure and interest organizations?
Although quite some studies have focused on the relationship between trust and formal social capital (e.g., Putnam, 2000; Uslaner and Brown, 2005: participation in political and civic life), research that looks at the relationship between trust and donating behavior—as a particular aspect of formal social capital—is rather scarce. Uslaner and Brown (2005) studied the relationship between generalized trust and donating to charitable organizations for the USA at the state level and found that the higher the aggregate level of generalized trust is within a state, the higher the percentage of people donating to charitable organizations. For the Netherlands, Bekkers (2003) showed that trusting people donate more money to charitable organizations than less trusting people, whereas for the USA, Brown and Ferris (2007) showed that the higher one's social and interracial trust (combined in one measure), the more likely one is to give for secular causes.
There are several routes to advance upon the current state of the art described previously. In so far as the relationship between trust and donating is studied, generalized trust is used as the most important trust indicator. The implicit mechanism underlying the relationship is the expectation that people who generally have a positive view on (unknown) others donate more (often) because they experience a higher level of efficacy of their donation (Bekkers and Wiepking, 2011). Yet, institutional trust—that is, the extent of trust in governmental institutions such as the legal system, the parliament, and the police—might be equally, or maybe even more important for the judgment of people on how well their money will be spent. A first advancement therefore is to study institutional trust next to generalized trust.
A second way to make progress is to distinguish between several types or organizations one can donate to, such as activist, leisure, and interest organizations. These types of organizations differ in the goals they pursue and the interests they advocate (e.g., Van der Meer et al., 2009). Van der Meer and colleagues showed individual characteristics, such as gender and religiosity, to differentially affect being involved in these types of organizations, involvement being a combination of membership, active participation, volunteering, and donating. However, they did not address the differential impact of generalized and institutional trust. Yet, the impact of one's judgment of the efficacy of a donation might differ conditionally upon these differences in goals and interests.
A third way to advance is by studying individual donating behavior as being conditional on the national context in which individuals live. Uslaner and Brown (2005) used single-level macro analyses to conclude that a higher level of aggregate trust leads to more donations. It is, however, not unthinkable that this is not an effect of community trust, but of individual trust. Generally, it is difficult to draw conclusions about individual level actions and relationships in macro-level analyses (Robinson, 1950). We therefore propose to use multilevel analysis to predict individual level donating behavior by using aggregated institutional and generalized trust as national level characteristics, while accounting for other important national level characteristics as income inequality and wealth (e.g., Uslaner and Brown, 2005) and individual levels of trust. Using logistic multilevel regression analysis for 20 countries in the European Social Survey of 2002, we test hypotheses at the individual and contextual level.
Theory and hypotheses
While donating behavior has been the subject of extensive research (Bekkers and Wiepking, 2011), most of this research is not concerned with distinguishing between types of organizations to donate money to. Nevertheless, it is generally understood that associational life comprises a wide variety of different organizations in which citizens can be active, ranging from sports clubs, human rights groups, and religious organizations to trade unions and political parties. All these organizations pursue diverse goals and serve particular interests, but what they have in common is that donations are a dominant source of funding. On the basis of the primary aim that these organizations focus on, Van der Meer et al. (2009) have proposed—and in terms of cross-national equivalence empirically validated—a typology of three types of voluntary associations. Activist organizations primarily advocate broad, social interests (e.g., humanitarian and environmental issues), leisure organizations fulfill recreational purposes (e.g., sports and culture), and interest organizations aim to defend their members' socio-economic interests (e.g., trade unions and consumer organizations). The authors put great emphasis on the importance of this distinction by showing that individual level characteristics differently affect involvement in these three types of organizations (Van der Meer et al., 2009). It is not unthinkable that contextual factors, such as national levels of generalized and institutional trust, have differential impacts on these types of organizations. Against this background, we set out to systematically investigate the effects of individual and national levels of generalized and institutional trust in order to explain why people donate to activist, leisure, and interest organizations. Moreover, we use the notion of efficacy—that is, the perception of donors that their contribution makes a difference to the cause they are supporting (Bekkers, 2003; Bekkers and Wiepking, 2011)—as the mechanism through which trust and donating are linked.1
Individual levels of generalized and institutional trust and donating
Generalized trust can be defined as trust in unknown others. The most frequently used measure of generalized trust in social science research is the survey question: “Generally speaking, do you believe that most people can be trusted or that you can't be too careful in dealing with people?” It is based on a worldview that stresses optimism and a sense of control (Uslaner and Brown, 2005). Uslaner (2008) states that trusting people are more likely to have positive beliefs about human nature and have faith in the good intentions of others, making them more likely to connect with people through good deeds, such as volunteering and donating. This claim is supported by empirical evidence. Bekkers (2003) demonstrated that the level of generalized trust is positively related to the amount of money donated to charity in the Netherlands. Brown and Ferris (2007) found such a positive relationship for the USA. Similarly, generalized trust may have a positive effect on people's willingness to donate to several types of organizations. If people believe that most people can be trusted and have faith in the good intentions of others, then this belief is likely to influence donations to all three types of organizations, regardless of their goals.
A particularly powerful explanation as to why individual generalized trust is positively related to donating behavior can be found in the notion of efficacy: the “perception of donors that their contribution makes a difference to the cause they are supporting” (Bekkers and Wiepking, 2011: 943). For donors, it is often unclear how their money is actually spent and it is this lack of transparency that causes potential donors not to give; it triggers the feeling that their effort does not matter anyhow (Bekkers, 2003). It has been shown that when people perceive their gift to make less of a contribution, their willingness to give diminishes (e.g., Smith and McSweeney, 2007; Wiepking et al., 2010). We expect individual generalized trust to have a strong impact on perceived efficacy. Through their general belief in “the goodness of man” for trusting people, the lack of transparency is less of an obstacle, because they believe—at least more so than less trusting people—that their donation will be used for the supposed cause. We thus hypothesize (Hypothesis 1) that the higher an individual's level of generalized trust, the more likely he or she is to donate to (a) activist, (b) leisure, or (c) interest organizations.
It is, however, possible that the strength of the effect varies across different types of organizations. In the case of activist organizations for example, the beneficiaries are usually unknown others, and donors are frequently passively involved (so-called “checkbook members”), whereas donors to leisure organizations are often active in these organizations themselves, and the receivers of the donated amount of money spend it on causes that are more likely to be visible to the donor. As a result, the effect of generalized trust may be strongest for activist organizations, where the transparency is lowest, and weakest for leisure organizations, where transparency is highest.
Whereas generalized trust is concerned with confidence in unknown others, institutional trust refers to confidence in governmental institutions, such as the parliament and the police, but also in organizations as such. Recent studies in philanthropy have emphasized the vital importance of confidence in charitable organizations for making donations (see, e.g., Bennett, 2003; Bekkers, 2003, 2006). Donors, who have more confidence in institutions and organizations, believe that their money is spent on the cause they are supporting rather than on excessive fundraising and overhead costs. As argued previously, such beliefs about the efficacy of institutions and organizations are likely to encourage donating behavior (Bekkers and Wiepking, 2011). On the basis of US data, the Independent Sector (2002) reported a difference of 50% in the annual amount of money donated to charities between people with a high and a low level of institutional trust. Similarly, Bekkers (2003 and 2006) showed that donor confidence increases the amount of donations to charitable organizations in the Netherlands. Additional to generalized trust, institutional trust might thus have an independent effect on donating, through the mechanism of perceived efficacy. In effect, because of the fact that it pertains to institutions and organizations and not to an abstract notion such as “mankind in general,” its impact might even be more pronounced than that for generalized trust. We hypothesize (Hypothesis 2) that the higher an individual's level of institutional trust, the more likely he or she is to donate to (a) activist, (b) leisure, or (c) interest organizations.
In addition, we examine the possibility that this relationship is stronger for activist than for leisure and interest organizations. The reason is that trust is more important in social interactions with higher uncertainties. When the risk of a certain action is higher, people need higher levels of trust to engage in that action (Coleman, 1990). For different types of voluntary organizations, the degree of uncertainty accompanied by their activities varies greatly. Wiepking (2010) found that when charitable organizations have goals that are carried out far away and are more difficult to accomplish, such as the Millennium Development Goals, the uncertainty that the donation will benefit the recipients is higher, making trust more important. By contrast, for charitable organizations that are more locally oriented and have goals that are easier to reach, the level of confidence in organizations did not affect the incidence of giving (Wiepking, 2010). Once again, the level of perceived efficacy is important for understanding the link between, in this case, institutional trust and donating behavior.
We thus expect that the effect of institutional trust is weaker for leisure and interest organizations as, in these examples, the degree of uncertainty is lower. Leisure organizations mainly offer socializing and recreational activities and are usually characterized by high levels of active participation and face-to-face contact (Van der Meer et al., 2009). As a result of this proximity, it becomes easier for people to monitor their organization's activities. Interest organizations, on the other hand, primarily defend their own members' socio-economic interests. People are more likely to donate money when they feel that their interests are being advocated. Even though active participation and face-to-face contacts are less prevalent in this type of organization as compared with leisure organizations, the goals are more easily tied to individual interests as compared with activist organizations.
National levels of generalized and institutional trust and donating
Whereas we expect that at the individual level generalized and institutional trust are related to donating in a similar vein, at the national level their impact might not be in the same direction. In countries with high levels of generalized trust, we expect people to be more stimulated to donate because in these countries all people are likely to be affected by and to internalize general norms that encourage this kind of behavior. In their classic study Almond and Verba (1963) stated that “in the United States and Britain, the belief that people are generally cooperative, trustworthy and helpful is frequent […] and is directly related to one's propensity to join with others in political activity.” This most probably is also the case for donating to charitable organizations. In nations in which the general belief prevails that others are willing to act in accordance with collective goals, even less trusting people are probably willing to donate to charitable organizations. Yet in the opposite case of a low aggregate level of generalized trust, even individuals that generally do trust others might think twice. Uslaner and Brown (2005) studied the relationship between generalized trust and donating to charitable organizations for the USA at the state level. They found that the higher the aggregate level of generalized trust is within a state, the higher the percentage of people donating to charity. We hypothesize that (Hypothesis 3) the higher a country's level of generalized trust, the more likely one is to donate to (a) activist, (b) leisure, or (c) interest organizations. Most likely, the leading mechanism is that a stronger general belief in “the good of mankind” results in a higher average level of perceived efficacy, independent of one's individual level of generalized and institutional trust.
However, while institutional trust at the individual level is vital for making donations to voluntary organizations, at the macro-level we expect that a general belief of institutions functioning and performing well leads to a lower likelihood of donating to any type of organization. In countries with high levels of institutional trust, people may on average be confident that institutions take over certain responsibilities, such as caring for others. If generally one believes that enough financial resources are present to ensure a flourishing civic life and that this money is actually spent well, individuals might be less willing to donate, because they are certain that the actions that the government undertakes are enough. Scheepers and Te Grotenhuis (2005) have actually shown that in stronger developed welfare states, citizens were less willing to “alleviate the poor,” suggesting that in these states individuals believe in power of the government to tackle these social ills, making their own contributions less necessary. Similarly, we hypothesize that (Hypothesis 4) the higher a country's level of institutional trust, the less likely one is to donate to (a) activist, (b) leisure, or (c) interest organizations.
Data: the European Social Survey
The European Social Survey is well suited for our purposes. This cross-national data source uses probability procedures in order to maximize the likelihood of the samples being representative. Response rates are high as a result of the organization's target to reach response rates of at least 70% (see www.europeansocialsurvey.org for fieldwork documentation). Besides being a high quality, cross-nationally comparable data source, the 2002 edition contains an extended measure of donating to various types of voluntary associations and of generalized and institutional trust. Moreover, this data source contains age, gender, educational level, urbanization, marital status, and church attendance as control variables for the relationship between trust and donating. After selecting 18–80-year-old respondents and listwise deletion, our analytic sample contains 33 474 respondents located in 20 countries (19 European countries and the USA).
Measurements: dependent and independent variables
We constructed three dependent variables: donations to activist organizations, donations to interest organizations, and donations to leisure organizations (e.g., Van der Meer et al., 2009). During the interview, the respondents were shown a list of organizations, for which they could mark whether they donated to them in the past 12 months. We used donations to humanitarian and to environmental, peace, and animal organizations as indicators of activist organizations, donations to trade unions, to business, profession, and farmer organizations, and to consumer and automobile organizations as indicators of interest organizations, and to sports and outdoor clubs, to cultural and hobby organizations, and donations to social clubs as indicators of leisure organizations. The three dependent variables are dichotomies, code 1 exemplifying that the respondent donated to at least one of the organizations belonging to that specific type.
Our central independent variables are individual level generalized trust, individual level institutional trust, national level generalized trust, and national level institutional trust. The three items “most people can be trusted or you can't be too careful,” “most people try to take advantage of you, or try to be fair,” and “most of the time people are looking after themselves or try to be helpful” were designed to measure generalized trust. Respondents could answer varying from 0 “you can't be too careful” to 10 “most people can be trusted”. There were five items pertaining to institutional trust: to what extent do you have trust in “your country's parliament,” “your country's legal system,” “the police,” “politicians,” and “the United Nations.” Respondents could answer from 0 “no trust at all” to 10 “complete trust”. We first constructed scales for both types of trust at the individual level. A principal factor analyses including all items showed two distinct dimensions: one pertaining to generalized trust, and the other to institutional trust (eigenvalues were 1,38 and 3,66, respectively). All items provided a strong contribution to the factor solution, with high commonalities and strong loadings. The dimensions of trust correlated with 0,51: the more one trusts one's fellow citizens, the more one trusts (inter)national institutions. We decided to calculate the mean score on the items belonging to the separate dimensions, so that the final scales range from 0 (no trust) to 10 (complete trust). Reliability analysis shows individual level generalized trust is measured very reliably (α = 0.77), as is individual level institutional trust (α = 0.83). To measure both types of trust at the contextual level, we aggregated the individual scores to the level of the country.
As control variables at the individual level, we include age (18–80 years), gender (male, female), educational level (primary, secondary, and tertiary educated), urbanization (large city, small city, rural area), income (standardized within countries on a scale of 0–100), marital status (married, single, divorced, widowed), and church attendance (never, scarcely, monthly, weekly). As income inequality (Gini) and national wealth (relative gross domestic product per capita, the European Union total in 2002 being fixed at 100 as a reference) have shown to be important common causes of both national levels of trust and social capital (Schofer and Fourcade-Gourninchas, 2001; Uslaner and Brown, 2005), we include them into our statistical multilevel models as national level control variables. Both variables are provided by Eurostat. For the analyses, all contextual characteristics are centered at their grand means. Table 1 contains the descriptive information of all variables.
Table 1. Descriptive statistics at the individual and contextual level
|Gender (female = reference)||0.00||1.00||0.52|| |
|Large city (reference)||0.00||1.00||0.34|| |
|Small city||0.00||1.00||0.59|| |
|Rural area||0.00||1.00||0.07|| |
|Married (reference)||0.00||1.00||0.58|| |
|Not religious/never||0.00||1.00||0.30|| |
|Gross domestic product per capita||48.00||240.00||116.33||35.62|
In Table 2 we present, first, the incidence of donating to the three types of organizations and, second, the average level of generalized and institutional trust. To facilitate interpretation, we sorted the countries on the basis of the proportion (highest to lowest) of people donating to activist organizations. Swedish and Dutch citizens donate most often to activist organizations; in these countries, one-third of the populations donated to such types of organizations in the past 12 months. In Greece and Hungary merely 2% of the population donated to activist organizations. Donating to interest organizations is most widespread in Slovenia, where 14% donated in the previous year, whereas in Hungary and Italy, one out of a hundred persons donated. Donations to leisure organizations occurred most often in Norway and Sweden, where almost one-fifth of the population gave money. In Hungary, the lowest level of giving is measured, with an incidence of 2%. Clearly, there are substantial differences between countries in the incidence donating to these three types of organizations.
Table 2. Percentage of people that donates to different types of organizations and of generalized and institutional trust by country
Average levels of generalized and institutional trust are highest in Denmark, Finland, Norway, and Sweden, where averages are six or higher (on a scale of 0–10). In Greece and Poland, the average of generalized trust is lowest, with 3.42 and 3.77, respectively, whereas in Poland and Slovenia institutional trust is lowest (more or less four). Here also, there seem to be substantial country differences, making it worthwhile to assess the influence of the national levels of trust on the average levels of donating to activist, interest, and leisure organizations.
Multivariate results: hypothesis testing
As our three dependent variables are binary and individuals are nested within countries, we estimated two-level random intercept logistic regression models. Table 3 first shows, per dependent variable, the null model. After that we will subsequently include individual level generalized and institutional trust and the individual level control variables (Model 1), and national level generalized and institutional trust and the contextual level control variables (Model 2). Only the effects of our central independent variables will be discussed.
Table 3. Donations to activist, interest, and leisure organizations regressed on individual and contextual level trust, multilevel random intercept models, 33 474 individuals, and 20 countries
|Age|| ||0.005**||0.006**|| ||−0.000||0.000|| ||0.010**||0.010**|
|Male (reference)|| || || || || || || || || |
|Female|| ||0.337**||0.349**|| ||−0.554**||−0.558**|| ||−0.415**||−0.421**|
|Income|| ||0.006**||0.006**|| ||0.011**||0.011**|| ||0.009**||0.009**|
|Married (reference)|| || || || || || || || || |
|Single|| ||0.059*||0.062*|| ||−0.373**||−0.376**|| ||−0.106**||−0.108**|
|Divorced/separated|| ||0.058||0.060|| ||0.060||0.058|| ||0.005||0.003|
|Widowed|| ||−0.096*||−0.098*|| ||−0.553**||−0.555**|| ||0.032||0.032|
|Large city (ref)|| || || || || || || || || |
|Small city|| ||−0.057*||−0.062*|| ||−0.039||−0.041|| ||0.170**||0.169**|
|Rural area|| ||−0.193**||−0.199**|| ||0.326**||0.329**|| ||0.317**||0.313**|
|Primary educated (reference)|| || || || || || || || || |
|Secondary educated|| ||0.698**||0.740**|| ||0.432**||0.438**|| ||0.405**||0.409**|
|Tertiary educated|| ||1.267**||1.330**|| ||0.955**||0.963**|| ||0.714**||0.722**|
|Never attends church (reference)|| || || || || || || || || |
|Scarcely attends church|| ||0.126**||0.129**|| ||0.145**||0.147**|| ||0.279**||0.284**|
|Monthly attends church|| ||0.289**||0.301**|| ||0.288**||0.290**|| ||0.497**||0.509**|
|Weekly attends church|| ||0.242**||0.248**|| ||0.183**||0.180**|| ||0.293**||0.296**|
|Generalized trust|| ||0.068**||0.070**|| ||0.022*||0.022*|| ||0.059**||0.058**|
|Institutional trust|| ||0.053**||0.054**|| ||0.017||0.019|| ||0.027**||0.027**|
|Generalized trust|| || ||1.118**|| || ||0.235|| || ||0.731**|
|Institutional trust|| || ||−1.052**|| || ||−0.888**|| || ||−0.798**|
|Income inequality|| || ||−0.054*|| || ||0.009|| || ||−0.005|
|Gross domestic product per capita|| || ||0.006**|| || ||0.007*|| || ||0.005*|
|Level 2 variance||0.526**||0.448**||0.199**||0.548**||0.569**||0.417**||0.329**||0.282**||0.178**|
|Intra class correlation||0.138||0.120||0.057||0.143||0.147||0.112||0.090||0.079||0.051|
The null models underscore the picture that arose from the descriptive findings: there is ample variation between countries in the average incidence of donating to the three types of organizations. All country level variance components are significant and the intra class correlations quite substantial: 0.139, 0.153 and 0.100 for donating to activist, interest, and leisure organizations, respectively, meaning that, respectively, 13.9%, 15.3%, and 10.0% of all individual differences in donating behavior can be attributed to the country in which they live.
Looking at Models 1, the effects of generalized trust show that with each increase of one unit on the scale of 0–10, the odds of donating versus not donating to activist organizations increase with 7.0% ([e0.068 − 1] ∗ 100), to interest organizations with 2.2% (b = 0.022), and to leisure organizations with 6.1% (b = 0.059). As all coefficients are significant, hypothesis 1 is supported. Furthermore, the effect of individual generalized trust is strongest for donating to activist organizations. This is what we expected, because of the relatively low level of transparency about how well the donated money is spent within those organizations. Unexpectedly however, the effect for leisure organizations is stronger as compared with the effect for interest organizations, even though we assumed that transparency is highest for the former. Hypothesis 2 on the influence of institutional trust is partly supported. One unit increase in institutional trust results in a 5.4% higher odds to donate to activist organizations as compared to refraining from donating (b = 0.053, p < 0.01). For leisure organizations, a one unit increase results in a 2.7% higher odds (b = 0.027, with p < 0.01). For donating to interest organization, we did not find a significant influence of individual institutional trust. Interestingly, the order of the strength of the effects of institutional trust is similar to the order of the effects of generalized trust (activist > leisure > interest). Unexpectedly however, the effect of individual institutional trust is less strong than the effect of individual generalized trust.
Considering Models 2, we find that the higher a nation's aggregated level of generalized trust, the higher the average incidence of donating to activist organizations (b = 1.118, p < 0.01) and to leisure organizations (b = 0.731, p < 0.001), but not to interest organizations (b = 0.235, p > 0.05). These results partly confirm our third hypothesis. Our fourth hypothesis is fully supported by our models: the higher the national level of institutional trust, the lower the percentage of people that donates to activist organizations (b = −1.052, p < 0.001), to interest organizations (b = −0.888, p < 0.001), and to leisure organizations (b = −0.798, p < 0.001).
Conclusions and discussion
In this paper, we answered the question as to what extent donating to activist, interest, and leisure organizations is affected by individual and national levels of generalized and institutional trust. On the basis of the notion of efficacy—the perception of donors how much their contribution will make a difference to the cause they are supporting—we deduced hypotheses on the relationship between trust and donating. We used the European Social Survey 2002 to estimate multilevel random intercept models, on the basis of more than 33 000 individuals living in 19 European countries and the USA.
Our results reveal that, firstly, there is ample cross-national variation in the average incidence of donating to these three types of organizations. Secondly, at the individual level, both generalized trust and institutional trust proved to positively relate with donating to all types of organizations (except for the impact of institutional trust on interest organizations). The effects are strongest for donating to activist organizations. This was expected, as for these types of donations, it is the least transparent how well the money of donors is spent and how well the goals of the organizations are met. In those circumstances, high levels of both generalized and institutional trust are vital for an individual's decision to donate. Thirdly, at the national level, generalized trust positively affects donations to activist and leisure organizations, but not to interest organizations. Furthermore, national level institutional trust negatively affects donations to all types of organizations.
The opportunity to incorporate important control variables such as educational attainment, income, church attendance, marital status, and urbanization strengthens our conclusions in terms of unbiased effects of an individual's generalized and institutional trust. Taking account at the national level for income inequality and wealth—which in the literature have been identified as the important common causes of macro-level trust and civic engagement (Uslaner and Brown, 2005)—strengthens our belief in the effects of national levels of trust not being severely biased. As such the findings contribute substantially to the existing literature on the explanations of giving behavior. However, because of possible omitted variable bias—we were unable to include individual values as a common cause of trust and giving—and not having estimated fixed-effects models, there might be some bias left in the effects of individual and national level trust. Future research might study the extent to which this is the case.
Clearly, the results for donating to activist and leisure organizations are much similar whereas for donating to interest organizations, the results are quite dissimilar. This by itself justifies the distinction into three different types of organizations to which individuals can donate. Why can it be the case that at the individual level, institutional trust is not related with donating to this type of organization (for the other two types of organizations the effect is present at the individual level), whereas at the national level, aggregated generalized trust does not affect it (for the other two types of organizations the effect is present at the national level)? In line with Van der Meer et al. (2009), we used trade unions, business, profession, farmer, consumer, and automobile organizations to indicate donations to interest organizations. As the four first mentioned all pertain to one's labor market situation and donations to and memberships of these organizations might be a matter of custom, rather than the outcome of an individual process of decision making, an individual's trust in (inter)national institutions might not be that important. Probably trust is not needed to donate to interest organizations, as people have “contract” type of agreements with these organizations (e.g., Lyons, 1994). The question is what we would have found if measures on trust in these specific organizations were present. Interestingly however, at the national level institutional trust does matter. This might indicate that in countries where on average one believes (inter)national institutions to perform well, the general idea is widespread that the government takes care of many of the interest of many citizens, making donations to interest organizations less necessary. The reason why national level generalized trust does not affect donating to interest organizations over and above one's individual level of generalized trust, might be found in these organizations advocating very specific interests, making a general belief in the cooperativeness and trustworthiness of others (Almond and Verba, 1963) unimportant, once individual level differences in generalized trust have been accounted for.
Generally, increasing individual levels of trust, either generalized or institutional, positively affects donating to almost all types of, but particularly to activist organizations. Clearly, for the purpose of raising more money for alleviating general social ills not directly tied to individuals, and for goals that actually do improve one's individual situation, such as better leisure organizations, increasing individual and average levels of generalized trust, as well as individual levels of institutional trust is a sound route to take. There is a large body of literature that uncovers the individual and national level determinants of civic engagement or trust. At the individual level, particularly educational attainment increases one's trust in unknown others and civic engagement in general (e.g., Gesthuizen and Scheepers, 2010; Gesthuizen et al., 2008; Helliwell and Putnam, 2007). If the goal is to increase generalized trust to achieve a higher willingness to donate, investing in the education of citizens is a sound strategy. Meanwhile, there might be ways of increasing the adequacy of reaching the most trusting people, that is, those having the highest levels of education and income (see results), as they have a higher propensity to donate. Higher education institutions and large professional organizations in high segments of the labor market might be willing to provide space to reach people for the purpose of donating to charitable organizations, particularly if fundraisers would be able to obtain access through persons who work for the organization.
The story is different, however, for national levels of institutional trust: the higher the average level of institutional trust of a country, the fewer citizens donate to any type of organization. As such, this is a positive finding: it suggests that when institutions are perceived to function well, individuals estimate that their charity is less needed. This situation is not negative if governments have enough finances—for instance through taxes—to spend on civic life in general. But if this is not the case and institutions are judged to be trustworthy, the risk increases that there is a lack of resources to maintain a healthy civil society. And obviously, when the functioning of civil society is at stake, this would be an undesirable situation.
We are, however, unable to empirically test this mechanism as in our data set a measure of efficacy is lacking.
Anouk Evers holds a Bachelor's degree in Cultural Anthropology and Development Studies and is currently a Research Master student of Social and Cultural Sciences at the Faculty of Social Sciences, Radboud University Nijmegen, the Netherlands. Her research interests include poverty, gender, and health issues in developing countries.
Maurice Gesthuizen is Assistant Professor at the Department of Sociology, Radboud University Nijmegen, the Netherlands. He studies causes and consequences of educational inequalities and of economic vulnerability and causes of a lack of social capital. Recently, he has published on these issues in journals such as Social Science Research and Sociology of Health and Illness.