Trust and institutions: A multilevel analysis

https://doi.org/10.1016/j.socec.2011.04.015Get rights and content

Abstract

In this study, we argue that levels of trust are shaped by formal and informal institutions. We discuss statistical test results that support this view. These utilize World Value Survey results and related data sources to examine the ways that institutions and contextual factors affect individuals’ self-reported trust levels. This investigation is the first, to our knowledge, to take into account micro-unit along with macro-unit characteristics to analyze the nature of contextual effects and does so by applying a multilevel regression approach.

Highlights

► We clarify the meaning and the “multilevel” mechanisms of the trust. ► The multilevel regression method is used to analyze these mechanisms. ► The results indicate that contextual effect matters. The trust can be promoted by higher quality of legal systems. ► The individual-level and group-level effects of the religious denominations on trust are different. ► And the relationship between the individual-level predictors and trust varies significantly from country to country.

Introduction

Trust is essential to economic interactions, success and plenitude. In primitive societies, people relied on a small band of fellows for survival. Today we voluntarily and routinely extend trust in order to engage in numerous interactions with large numbers of strangers so that we can have access to specialized goods and services. For example, we risk our hard-earned money to prepay someone whom we have never met for all sorts of items purchased over the Internet. We seek large numbers of exchange opportunities and we expect that there are not just formal rules but also informal norms of behavior that accompany most transactions and constrain opportunistic temptations, or even punish those trading partners who fail to deliver goods or services in good faith. The sense of security and confident reliance that we have in most of our transactions is built on formal as well as informal “rules of the game” that punish the defectors for breaches of trust. In this sense, trust is wound up with the institutions that characterize successful market economies. Many now refer to it as a foundation for a civil society.

Nevertheless, the concept of “trust” has a relatively “broad array of meanings and connotations” in the research literature and in everyday use with a variety of definitions (Mcknight and Chervany, 1996, p. 6). For example, in the available literature, trust has been viewed as a concept which retains many of the meanings employed in everyday language, such as the “expectancies”, “confidence”, “beliefs”, “affective”, “cognitive”, “competence”, “benevolence”, “dependence”, “reliance”, “security”, etc. (Husted, 1990, Blakeney, 1986, Giffin, 1967, Williamson, 1993, Mcknight and Chervany, 1996, Shapiro, 1987). These researchers define “trust” to encompass its psychological, “cognitive, emotional, or behavioral dimensions” which can be merged into different social interactions (Lewis and Weigert, 1985, p. 969). Because of the multi-dimensional construction of trust, the definitions of it in the literature tend to reflect various specific yet narrow domains of trust. For example, some scholars regard trust as a personal or inter-personal attribute embedded in human interactions, and promoted by personal bonds, community rules, informal mechanisms, or recurrent transactions, etc. (e.g., Granovetter, 1985, Erikson, 1968, Rotter, 1967, Geertz, 1978, Macaulay, 1963). Formal institutions in their view, are taken as a control or deterrent device rather than a form of trust support because they are only used when trust breaks down (Shapiro, 1987). However, other scholars refer to trust as an institutional property or regard it is fundamentally shaped by institutions, especially by formal institutions and general social similarities (Shapiro, 1987, Zucker, 1985, Yamagishi and Yamagishi, 1994, Yamagishi et al., 1998). They argue that personal bonds, intensive group ties, or local mechanisms may work well to generate a sense of security or trust within very limited social boundaries, either in the form of tribes, families, or local communities, etc. But in less personal, faceless or large scale exchanges that are common in modern market economies, social control measures derived from social ties and direct personal contact can be very limited and fragile. Yet they are relied on to produce trust in a disenchanted world, formal institutions can “serve as a springboard” to promote or constrain cross-situational trust on its own account (Gambetta, 1990, Zucker, 1985). In a study by Pearce et al. (2000), these writers even “see a minimum level of institutional trust as sine qua non for the emergence of interpersonal trust” (Pearce et al., 2000, Rousseau et al., 1998, p. 401, italics are authors). Another way to define “trust” is to view it as a dispositional attribute of mankind or “a general outlook on human nature” (Uslaner, 2004, Wrightsman, 1991). You trust because you have a general faith in human nature. From this human nature point of view, trust can be “conceptualized as a cross-situational, cross-personal construct” (Mcknight and Chervany, 1996, p. 37). This construct recognizes that an individual “has a consistent tendency to trust across a broad spectrum of situations and persons” (Mcknight and Chervany, 1996, p. 37, italics are authors). Meanwhile, there are various other scholars who use game theory to exam trust through cooperation or competition activities (Axelrod, 1984; Deutsch, 1960, 1962). Trust in these studies is then conceptualized as an action.

In this paper, we define trust as an “intention” rather than an “action”. It refers to a different degree of intention or willingness from one party to expect another party (parties) to act competently and dutifully “in a risky course of action” (Lewis and Weigert, 1985, p. 971). This trusting intention is an individual-specific and situation-specific concept. It is a property or state of a person directed to another person or persons. We may build our trust based upon how we assess the probabilities of gain and loss, or how we calculate expected utility by using “hard performance data” to conclude whether the person in question would “behave in a predictable manner”.2 In this sense, trust has more cognitive and calculus aspects. But in some other situations, emotional elements can take a dominant position in the formation of trust. For instance, affective bonds of love between family members or friends may create a social situation in which unconditional trust is promoted. In terms of the attributes of the trusted person or persons that the trustor evaluates, they can vary from situation to situation. For example, when a client trusts a bank teller, it is more likely that she believes the bank teller is capable of helping her and will not cheat her. Her trusting intention depends more on the teller's capabilities or bank's reputation rather than on expectations of the goodness in the teller's heart. By contrast, trust from children to parents is more linked to the love from the parents rather than the achievements of the parents. Trust, in our view, is thus “a mix of feelings and rational thinking” (Lewis and Weigert, 1985, p. 972). The institutional, dispositional, “cognitive, affective, and behavioral contents of trust are present in every instance of trust to some extent”, though “their qualitative mix across instances of trust” may differ from person to person, or from situation to situation (Lewis and Weigert, 1985, p. 972). And it is these different compositions that provide a basis for distinguishing different types of the trusting intentions, such as general trust, calculative trust, emotional trust, interpersonal trust, etc., defined in the existing literature.

We also agree with the argument made by Zucker (1985) that one type of trust can be substituted for another if it becomes advantageous to do so. For instance, an increase in the price to produce one type of trust can result in an increase in demand for its substitute (Zucker, 1985, p. 18). Any type of the trust, just like altruism or sympathy, is a scarce resource (Gambetta, 1990, p. 8). And it is manufactured by individuals using different mechanisms. These mechanisms include “the veil of ignorance” and “original position”, which can set up the human nature component of trust (Rawls, 2005, Zucker, 1985, p. 19). But the intangibility and subtleties embedded in this natural disposition cannot easily be quantified or tested via empirical research. The second group of mechanisms is exemplified by formal macro-level institutions and general social similarities, such as law, regulations, contracts, religions, etc. Socialization, gift giving, or community rules which are effective in the arm's length transactions, etc., represent the third group of mechanism to produce trust.

It is important to note that when speaking of mechanisms, we are taking trust as the output or effect rather than a cause (Zucker, 1985, p. 10). We recognize that in some other studies, trust is treated as a cause to reduce transaction costs or to promote cooperation in the classic Prisoner's Dilemma or in various other human interactions (Williamson, 1975, Axelrod, 1984). However, trust in these studies usually takes a behavioral realization. It refers to a trust-implying action. “When we see others acting in ways that imply that they trust us, we become more disposed to reciprocate by trusting in them” (Lewis and Weigert, 1985, p. 971). In this paper, we exclude trust as a behavioral realization and only focus on trusting intentions. It may be argued that trusting or distrusting intentions may act as a cause to change micro-level social norms or community rules, just as in the actual cases studied by Ostrom (1990) or Ellickson (1991). But for macro-level or nation-wide institutions, how much of trusting or distrusting intentions from the citizenry would it take to undermine the banking system, or to incite revolutionary change in the political system? Answering these questions is beyond the scope of this paper. (Lewis and Weigert, 1985, p. 981).

Based on the available cross-sectional data, we therefore argue that nation-wide institutions3 have a causal effect on trust, not the other way around. Institutions, either formal or informal “rules of the game”, refer to laws, regulations, bureaucratic structures, or religions in this study. Our investigation will follow this causality logic to study how and to what degree institutions affect trust – after controlling the relevant socioeconomic factors. And accepting the fact that the idea of institutions is still regarded as a somewhat vague concept, trust might be one important way to understand the quality of institutions as it can be thought of as a mental picture or mental reflection of them.

Section snippets

Literature review

Meanwhile, we also place emphasis on the multilevel perspectives of the trust in this paper. As pointed out by Sitkin et al. (1998), “trust require(s) a multilevel analysis”, and “scholars across a variety of disciplines have embraced this multilevel view” (Sitkin et al., 1998, pp. 397–398). However, some of the existing literature conceptualized the “multilevel” characteristics of the trust as the level of analysis to reflect the array of individuals, groups, firms, cooperations “in which

Data

In this paper, we use various data sources to examine whether or to what degree individual trust is affected by micro-level as well as macro-level mechanisms. The dependent variable is “Trust”. For this variable, World Value Survey (WVS hereafter) provides data on self-reported levels of trust by individuals who responded to surveys during the years 2000–2008. WVS is a widely used data source cited in many of the previous studies. We use this data source to derive a dummy dependent variable

Empirical test results and interpretations

The Regression 1 in Table 3 contains a single explanatory variable to indicate the basic partition of the variability in the data between individual-level and country-level effects. According to this model, the total variance of reported trust can be decomposed as the sum of the individual-level and country-level variances. The estimates of individual-level variance (0.998) and country-level variance (0.713) yield an intra-country correlation coefficient of 0.417 to represent the proportion of

Conclusions

It is a truism that levels of trust between individuals are profoundly important. Recent research corroborates the idea that there are identifiable variations in levels of self-reported trust from individual to individual as well as from place to place and culture to culture. Yet this is not enough. Trust is, by definition, other-directed and we must know more about how the context within which the individual finds himself matters as we investigate the determinants of trust. This study is the

References (51)

  • J.W. Dawson

    Causality in the freedom-growth relationship

    European Journal of Political Economy

    (2003)
  • L. Guiso et al.

    People's opium? Religion and economic attitudes

    Journal of Monetary Economics

    (2003)
  • L.S. Wrightsman

    Interpersonal trust and attitudes toward human nature

  • D. Acemoglu et al.

    The colonial origins of comparative development: an empirical investigation

    The American Economic Review

    (2001)
  • R. Axelrod

    The Evolution of Cooperation

    (1984)
  • N. Berggren et al.

    Free to trust: economic freedom and social capital

    Kyklos

    (2006)
  • Beugelsdijk, S., von Schaik, T., 2001. Social Capital and Regional Economic Growth. CentER Discussion Paper,...
  • S. Beugelsdijk

    A note on the theory and measurement of trust in explaining differences in economic growth

    Cambridge Journal of Economics

    (2006)
  • C. Bjornskov

    Determinants of generalized trust: a cross-country comparison

    Public Choice

    (2006)
  • R.N. Blakeney

    A transactional view of the role of trust in organizational communication

    Transactional Analysis

    (1986)
  • A.S. Bryk et al.

    Hierarchical Linear Models

    (1992)
  • Central Intelligence Agency

    The Worldfact Book

    (2010)
  • J. Delhey et al.

    Predicting cross-national levels of social trust: global pattern or nordic exceptionalism?

    European Sociological Review

    (2005)
  • R.C. Ellickson

    Order Without Law: How Neighbors Settle Disputes

    (1991)
  • E.H. Erikson

    Identity: Youth and Crisis

    (1968)
  • Fraser Institute

    Economic Freedom Index

    (2010)
  • F. Fukuyama

    Trust

    (1995)
  • D. Gambetta
  • C. Geertz

    The bazaar economy: information and search in peasant marketing

    American Economic Review

    (1978)
  • K. Giffin

    The contribution of studies of source credibility to a theory of interpersonal trust in the communication process

    Psychological Bulletin

    (1967)
  • E.L. Glaeser et al.

    Measuring trust

    The Quarterly Journal of Economics

    (2000)
  • M. Granovetter

    Economic action and social structure: the problem of embeddedness

    The American Journal of Sociology

    (1985)
  • L. Guiso et al.

    Does culture affect economic outcomes?

    Journal of Economic Perspectives

    (2006)
  • F.A. Hayek

    Law, Legislation and Liberty, Vol. 1: Rules and Order

    (1973)
  • B.W. Husted

    Trust in business relations: directions for empirical research

    Business & Professional Ethics

    (1990)
  • Cited by (50)

    • The gender gap in trust in Banks

      2023, Research in International Business and Finance
    • What Information Do Shoppers Share? The Effect of Personnel-, Retailer-, and Country-Trust on Willingness to Share Information

      2020, Journal of Retailing
      Citation Excerpt :

      Our multilevel modelling reflects the multilevel nature of trust and sheds light on the dynamic mechanism through which trusts at the micro- and macro-levels work in tandem to influence individuals’ privacy decisions. As such, our study also adds to the growing body of research on multilevel trust (Rousseau 2003; Fulmer and Gelfand 2012; Wang and Gordon 2011) within the retailing context. Finally, our results corroborate recent criticisms of the overreliance on PCs to understand privacy (Martin 2016; Nissenbaum 2011; Martin 2020) by demonstrating that PCs’ influence is highly conditioned by the contextual elements of a privacy decision, such as the type of information to share, with whom, and the relationships between the parties involved in the information sharing.

    • Individual trust and quality of regional government

      2022, Journal of Institutional Economics
    • Distrust of banks among the unbanked and banked

      2023, International Journal of Bank Marketing
    View all citing articles on Scopus

    We thank Cheng Hsiao, Yongheng Deng, Xingyou Zhang, Jiyoung Park, Huio Liu, Xudong An and Peng Fei for their valuable suggestions and help. We also thank anonymous referees for their valuable comments. The views expressed in this paper are those of the authors alone, and any mistakes are our own.

    1

    Tel.: +1 213 740 1467; fax: +1 213 740 6170.

    View full text