Trust in Government and Income Inequality
Political trust is an evaluative orientation of citizens toward their political system, or some part of it, based upon their normative expectations.
Income inequality is the extent to which income is distributed in an uneven manner among a population.
In the mid-1960s, Americans had confidence in their political institutions. However, between 1964 and 2017, the percentage of Americans who trusted that the government in Washington would do what was right “most of the time” or “just about always” fell from 76% to 20% (Pew Research Center 2017). A number of comparative studies examined whether other Western democracies experienced a similar reduction in trust of their political institutions. These studies revealed that declining trust in government was not unique to the United States. Indeed, diminished levels of trust in government were happening simultaneously in countries with different political systems and cultures (Houston et al. 2016). As trust in politicians and institutions decline, support for democracy did as well. Thus, declining trust is one of the central problems in modern politics. Another recurring topic that has emerged in both academic literature and popular culture is increasing income inequality. The literature suggests that the gap between the rich and the poor has been rising not only in the United States but also around the world. Income inequality is an important indicator of government performance. Studies show that high levels of income inequality decrease citizens’ trust in government across various countries around the world.
Definition of Political Trust
The concept of trust is both simple and complex. It is simple because it is used in daily language. Yet, it is inherently complex because it is used to explain a wide variety of social concepts. Houston and Harding (2013) define trust as “a willingness to rely on others to act on our behalf based on the belief that they possess the capacity to make effective decisions and take our interests into account” (p. 55). These scholars distinguish two dimensions of trust: competence and care (trustworthiness). The competence or cognitive dimension is based on rational judgments about the ability of the trustee to accomplish a stated goal and act consistently. On the other hand, trustworthiness is an effective dimension. It is grounded on the belief that the trustee’s action is not motivated by self-interest but rather takes into account the interests of the trustor. LaPorte and Metlay (1996) describe it as “the belief that those with whom you interact will take your interests into account, even in situations where you are not in a position to recognize, evaluate, and/or thwart a potentially negative course of action by ‘those trusted’” (p. 342).
Levi and Stoker (2000) suggest that a definition of trust entails elements described below. Trust is a belief about the future actions or inactions of others and their outcomes. Trust is relational: X trusts Y to do A, in doing so, X makes him/herself vulnerable to Y. Thus, trust involves risk since there is no certainty about the future outcomes. Trust is a judgment. One can trust or distrust the other, or trust or distrust to a certain degree. Trust judgments inspire a course of action. Distrust may inspire vigilance in and monitoring of a relationship. Thus, trust judgments reflect a fundamental belief about the trustworthiness of other people, groups, or institutions. The aforementioned definitions of trust entail to some extent interpersonal trust which is trust one person places in another individual. In contrast, political trust is the trust that people place in political institutions or their leaders.
Importance of the Trust
Why is trust so important? According to Putnam (2000), “Honesty and trust lubricate the inevitable frictions of social life” (p. 135). In many ways, trust acts as a facilitator for solving the shared collective problems of a pluralist society. Trust links citizens with the government and institutions that represent them therefore trust enhances both the legitimacy and the effectiveness of democratic government. No government can have the absolute trust of its citizens because the power of any government represents a threat to individual freedom and welfare. Notwithstanding this affirmation, government should have a minimum level of confidence of its citizens to operate effectively. Trust is fundamentally important for a healthy and functioning democracy (Mishler and Rose 2005). It engenders citizen compliance with public policies, encourages political participation, and contributes to perceptions of governmental legitimacy (Levi and Stoker 2000).
At stake is the efficacy and stability of democratic government, its ability to represent interests, and solve social and economic problems. Institutional trust translates into diffuse political support and is critical for “the survival and the effective functioning of democratic institutions” (Mishler and Rose 2005, p. 1051). Trust encourages the public’s acceptance of democratic values and ideals. To the contrary, political distrust undermines government legitimacy, threatens political stability, and facilitates support for undemocratic regimes (Mishler and Rose 2005). Further, trust stimulates citizen participation in political life (Uslaner 2002). Citizens who trust the government are more likely “to vote, follow politics, feel a sense of civic duty, and have high levels of political knowledge” (Mishler and Rose 2005, p. 1068). According to Gamson (1968), political mobilization and activism are driven by political trust. In the context of government, political trust influences whether or not citizens view the state as being politically legitimate, determines to what degree citizens will consent to government demands, and impacts the levels of political participation undertaken by citizens. Simply put, political trust is essential for the proper functioning of democratic governance.
In order for a democratic state to wield its authority in a productive fashion, it must be judged by its citizens as being legitimate. As Tyler (1998) declares, “Beyond being able to secure compliance, authorities need to be able to gain voluntary acceptance for most of their decisions. Legal, political, and organizational theorists have long recognized that voluntary acceptance of the decisions and rules of organizational authorities is important to the ability of those authorities to function effectively” (p. 271). As trust in politicians and institutions decline, so too could support for democracy itself. Without trust, citizens are less likely to pay taxes and support government policy (Yang and Holzer 2006).
Trust is especially important for generating public support for programs that entail some perceived risk or sacrifice (Hetherington 2005). Specifically, political trust is significant for the support of redistributive programs, such as welfare and food stamps, and race-targeted policies. When trust is high, the “haves” are more likely to make the sacrifices necessary to assist the “have-nots.” Hetherington observes that public policy making in the United States has become more conservative since the 1960s. However, public attitudes in American society have not become more conservative since this time. Hetherington explains this seeming contradiction by demonstrating a strong link between political trust, public opinion, and policy outcomes. Generally speaking, when the public is more trustful, the government responds with more liberal public policy. In contrast, when the public is more distrustful, the government responds with more conservative public policy.
At the same time, “a certain amount of rational distrust is necessary for political accountability in a participatory democracy” (Barber 1983, p. 166). In the extreme case, one can argue the public choice view of government as a Leviathan. Citizens should mistrust their government as this would be a natural position for them to defend themselves against a Leviathan. Distrust could be interpreted as a realistic view of government. However, this view fails to explain why the level of confidence was high in the past.
Declining Trust in Government
During the last several decades, many countries have moved from a nondemocratic government to a democratic one. However, it seems paradoxical that the geographical spread of democracy has been followed by “erosion” of its essential elements in old advanced democracies on both sides of the Atlantic. Crozier et al. (1975) define this phenomenon as a “crisis of democracy” in the nations of North America, Europe, and Japan. The main symptoms of weakening democracy are “erosion” of confidence in political institutions and the leaders of those institutions (Hetherington 2005). Decades of responses to the same survey questions show diminished political trust in government in a number of advanced industrial democracies (Levi and Stoker 2000).
The erosion of confidence has common traits across nations. First, the erosion of confidence is not a one-time phenomenon that is attached only to a particular economic or political event. The analysis of several surveys conducted since the 1960s has demonstrated that it is rather a persistent phenomenon that has been observed over the last three or four decades. For example, the decline in trust in the USA in the 1960s and 1970s was triggered by citizens’ reactions to the war in Vietnam, Watergate, and civil rights initiatives; however, the level of trust did not increase as the politicians associated with these events left political office (Levi and Stoker 2000). Such research findings suggest that declining trust is reflective of more than simply dissatisfaction with incumbents in office (Levi and Stoker 2000).
Second, the decline of public trust is an international phenomenon. It has been observed across the North American and European nations. Third, the decline of trust is structural in nature by permeating all political and social strata. Citizens lost confidence not only in political institutions but also in unions, big businesses, churches, televisions, and printed mass media as well as a decline in social capital (Putnam 2000). Lastly, the disenchantment is rather pragmatic and not of an ideological nature which has worsened with the economic difficulties in the nations.
Importance of Studying Income Inequality
There are two different types of distribution of income: absolute deprivation and relative income inequality. Absolute deprivation is the amount of income that the poor have relative to some measure of “minimally acceptable” income. For example, in the United States, the standard for measuring absolute deprivation is the poverty line. The relative income inequality is the amount of income that the poor have relative to rich. One can argue that once the poor reach a bare essential level of consumption, it should not matter how much money the rich have. Accordingly, why does income inequality matter? Should citizens care about how much the poor have relative to the rich? Does income inequality have a negative effect on trusting attitudes?
In example A, the distribution of voter preferences is relatively tightly clustered around the median. In example B, the distribution is much more spread out as compared to example A. Thus, two randomly chosen people from distribution B will be further from the median voter position, i.e., further from the selected policy. Therefore, dissatisfaction with government among citizens will be higher in example B. A similar argument could be applied to example C. The distribution in this case is bimodal. The median distance from the policy choice for a randomly chosen person is higher for example C as compared to example A. As the median distance from the policy choice increases, the level of trust in government decreases. Thus, the level of trust in government may be different in countries with different voter populations. In more homogeneous countries (example A), the level of trust is higher. In more heterogeneous countries with preferences distributed as in example B and example C, the level of trust is lower.
There has been an increase in the number of studies on the effects of societal heterogeneity on trust. Investigators have found that societal heterogeneity has a negative impact on trust: the more homogeneous a country is, the higher its trust, and vice versa. The explanation is that societal heterogeneity is conducive to the development of bonding trust in individuals of the same ethnicity, religion, and language at the expense of generalized (or bridging) trust. The previous research demonstrates that diverse societies are often challenged in generating and sharing public goods, trusting each other, and establishing well-functioning public institutions (Delhey and Newton 2005).
A homogenous society is more likely to induce a feeling of solidarity. Uslaner (2002) argues that “what matters is not how rich a country is, but how equitable is the dispersion of income” (p. 181; italics in original). Studies show that the level of inequality in society is negatively associated with measures of well-being. Luttmer (2005) finds that individuals’ self-reported happiness rises as their own income rises, but falls as the earnings of others in the area rise. Thus, relative income, and not absolute income, determines well-being.
There is ample theoretical and empirical evidence in the literature that wealthier citizens have different preferences when it comes to redistribution than poorer citizens and the wealthy have a disproportionate influence on government policies in democracies (with the implication that as inequality increases, the wealthy’s influence will increase as well). In addition, more affluent citizens have a broad understanding of the state of the economy and will hold the government accountable for the condition of the economy (Gilens 2005, 2012; Bartels 2009).
Multilevel Models for Cross-National Analysis
Multilevel models (also known as hierarchical models or mixed models) refer to models with a nested data structure (Hox 2010). Generally speaking, the individuals and the groups to which they belong are conceptualized as a hierarchical system of these individuals and groups, with individuals and groups each defined as a separate level of this hierarchical system. Thus, the sample data are viewed as a mutlistage sample from this hierarchical population. In educational research, for example, the population is schools and students nested within each school. Other examples include cross-national studies where the respondents are nested within their countries, or organizational research with workers nested within their firms.
Traditional regression methods only allow analyzing data on one level. Therefore, historically, multilevel data were either aggregated or disaggregated to one single level and then analyzed using either an ordinary multiple regression routine or any other “standard” statistical method. Aggregation involves moving variables at a lower level to a higher level, for example, by assigning to the countries the country mean of the respondents’ income levels. Disaggregation refers to moving variables to a lower level, for example, by assigning to all respondents in the countries an unemployment rate of the country they belong.
However, analyzing variables from different levels at one common level is inappropriate for two reasons. The first problem is statistical. First, aggregation leads to the loss of information and as a result to a loss of the statistical analysis power. On the other hand, disaggregation generates from a small number of country-level variables a lot of individual-level variables leading to an increase of the sample size of the data. However, the proper sample size for these data is a number of higher-level units, i.e., the number of countries. Ordinary statistical methods treat all these disaggregated data values as if all observations are independent from each other. As a result, statistical analyses produce “significant” results that are simply spurious (Hox 2010).
The second problem is conceptual. If researchers analyze data at one level but formulate conclusions at another level, they commit the fallacy of the wrong level. The ecological fallacy refers to interpreting aggregated data at the individual level. For example, Robinson (1950) finds that states with larger immigrant populations tend to have higher literacy. It is invalid to draw a conclusion that immigrants tend to be more literate than those born in the United States. In contrast, using relationships between variables on the level of the individuals to make conclusions about groups leads to the atomistic fallacy. For example, a study of individuals may find that an increase in income leads to an increase in trust. If it is inferred from these data that countries with higher gross domestic products have higher trust, then the atomistic fallacy is committed.
In the past, multilevel data were analyzed using conventional multiple regression analysis with one dependent variable at the lowest level and a set of independent variables from all available levels (e.g., Boyd and Iversen 1979; Van den Eeden and Huttner 1982). However, this approach leads to all of the statistical and conceptual problems discussed above. Standard statistical tests assume the independence of the observations. However, pooling hierarchically structured data into one sample violates this assumption. It is expected that two randomly selected individuals from the same group tend to be more alike than individuals selected from different groups. For instance, if a country experienced a major breach of public trust shortly prior to the administration of the survey, respondents in this country are likely to be watching the same news reports, reading the same newspaper account, reading the same editorials, and even talking with one another about the event. Thus, these respondent attitudes are not independent of one another as is assumed by OLS. Failing to account for this effect among respondents in distinct clusters (i.e., groups) results in OLS standard errors that are smaller than they should be and making it more likely that the null hypothesis of no relationship will be rejected. Hence, it increases the likelihood that a Type I error is made. In the extant literature, this effect of cluster sampling is known as a “design effect” (e.g., Kish 1987, 1995).
There are two ways to correct this lack of independence among respondents in a cluster (i.e., group unit). One is to report clustered standard errors. Cluster adjustment assumes that observations within groups are correlated but observations across groups are independent. These standard errors allow for a general form of heteroscedasticity but do not allow for errors to be correlated across or within groups (Huber 1967; White 1982). Thus, clustered standard errors account for both a general form of heteroscedasticity as well as for any intra-cluster correlation.
However, combining variables from different levels in one statistical model requires more complicated technique than estimating and correcting for the design effect. Multilevel models are designed to analyze variables from different levels simultaneously. It is the more sophisticated approach to addressing the issue caused by clustered cases. Multilevel models allow for simultaneous estimation of relationships at individual and country levels. They also allow for the relationships at the individual level to vary across countries. Furthermore, multilevel models permit the inclusion of cross-level interactions – i.e., examining how country-level factors influence individual-level relationships. In sum, multilevel modeling enables measuring the effect of different variables and explaining variation both within and across countries (Gelman and Hill 2007; Goldstein 2011; Raudenbush and Bryk 2010; Steenbergen and Jones 2002).
Dahl (1947) argues that comparison is what helps public administration achieve scientific status. Comparative studies of political attitudes are especially fruitful when they combine the particular political context in which people form those attitudes with critical individual-level variables. It leads to a more general model and comprehensive understanding of the forces that shape citizen political behavior. More and more comparative studies in public administration have started to examine politically and culturally diverse set of countries. Numerous studies have demonstrated and examined a declining level of trust in government and income inequality. Studies indicate that when inequality rises and trust in government falls, individuals may be less willing to support the maintenance and/or expansion of those features of government that have been found in the literature to reduce income inequality: public employment and certain types of welfare spending (Pontusson et al. 2002; Bradley 2003).
- Barber B (1983) The logic and limits of trust. Rutgers University Press, New BrunswickGoogle Scholar
- Bartels LM (2009) Unequal democracy: the political economy of the new gilded age. Princeton University Press, PrincetonGoogle Scholar
- Boyd LH, Iversen GR (1979) Contextual analysis: concepts and statistical techniques. Wadsworth Pub. Co., BelmontGoogle Scholar
- Crozier M, Huntington SP, Watanuki J (1975) The crisis of democracy: report on the Governability of Democracies of the Trilateral Commission. New York University Press, New YorkGoogle Scholar
- Gamson WA (1968) Power and discontent. Dorsey Press, HomewoodGoogle Scholar
- Gelman A, Hill J (2007) Data analysis using regression and multilevel/ hierarchical models. Cambridge University Press, CambridgeGoogle Scholar
- Gilens M (2012) Affluence and influence: economic inequality and political power in America. Princeton University Press, PrincetonGoogle Scholar
- Goldstein H (2011) Multilevel statistical models. Wiley, New YorkGoogle Scholar
- Hetherington MJ (2005) Why trust matters: declining political trust and the demise of American liberalism. Princeton University Press, PrincetonGoogle Scholar
- Hox JJ (2010) Multilevel analysis: techniques and applications. Routledge, New YorkGoogle Scholar
- Huber PJ (1967) The behavior of maximum likelihood estimates under nonstandard conditions. In: Le Cam LM, Neyman J (eds) Proceedings of the Fifth Berkeley symposium in mathematical statistics and probability. University of California Press, Berkeley pp 221–233Google Scholar
- Kish L (1995) Survey sampling. Wiley, New YorkGoogle Scholar
- Luttmer E (2005) Neighbors as negatives: relative earnings and well-being. Q J Econ 120(3):963–1002Google Scholar
- Pew Research Center (2017) Public trust in government remains near historic lows as Partisan Attitudes Shift. (Data File and Codebook). http://www.people-press.org/2017/05/03/public-trust-in-government-remains-near-historic-lows-as-partisan-attitudes-shift/
- Raudenbush SW, Bryk AS (2010) Hierarchical linear models: applications and data analysis methods. SAGE, Thousand OaksGoogle Scholar
- Tyler TR (1998) Trust and democratic governance. In: Braithwaite V, Levi M (eds) Trust and governance. Russell Sage Foundation, New York, pp 269–294Google Scholar