Perception of corruption as a business obstacle, generalized trust and relation centrism in low- and middle-income nations: the moderating influence of governance

Putnam (1993) theorized that trust, reciprocity and civic engagement are indispensable to collective existence and argued that communities become prosperous because they have a vital civic life. A multi-level characterization and complexity of trust has also been recognized in management studies (Rousseau et al. 1998; Zheng et al. 2023). Generalized trust is impersonal and not related to specific social exchange relationships between people (Ellwardt et al. 2012; Martinangeli et al. 2023). It is an abstract attitude towards the out-group and people in general, encompassing those beyond immediate familiarity, including strangers (people one randomly meets in the street, fellow citizens, and foreigners, amongst others) (Freitag and Traunmüller 2009; Zheng et al. 2023). Generalized trust is associated more with “weak ties” social capital that can create bridges, promote openness, exchange, knowledge diffusion, and decrease corruption (Uslaner 2004). These weak ties could make it easier to engage in corrupt practices because social capital investments may not be needed to engage in corrupt activities.

The social network may also encourage corruption through wide-ranging social pressures that influence the decisions of agents (Besser and Miller 2011; Danis et al. 2011). According to Hofstede’s individualism-collectivism cultural dimension (Hofstede 2011), people, in individualistic societies are expected to care primarily for themselves and their immediate families while in collectivist cultures, people view themselves as members of larger groups, including extended family members, and are expected to take responsibility in caring for such larger groups (Huff and Kelley 2005). Collectivistic societies are typically found in low- and middle-income countries of the Global South such as most of Latin America, Africa, Asia, and the Middle East, where bribe-taking is widespread (Sanyal 2005).

In this paper, we introduce a new concept called “relation centrism”, which we define as the importance that people give to family and friends in society. While generalized trust refers to an abstract trust in the out-group, relation centrism builds on the concept of in-group centrism to refer to trust in family members or friends (Bullough et al. 2017; Eckel et al. 2022; Kruglanski et al. 2006). Ethnic and non-ethnic ties affect firms (H. Li 2020; Santana et al. 2009; Yeung 1997; Zhu et al. 2022) and the literature has defined particularized relational trust as trust found in close social proximity and extended toward people the individual knows from everyday interactions (e.g., family members, friends, neighbors and co-workers) (Freitag and Traunmüller 2009; Zheng et al. 2023). Furthermore, social networks can enable the coordination of corrupt activities and override practices meant to govern firm behavior (Ju and Wang, 2023; Liu et al. 2024; Santana et al. 2009).

Particularized relational trust has been reported to facilitate corruption (Uslaner 2004), with the following logic: to form a bribery-corruption relationship (the transaction type corruption), some minimal trust must exist because of the time lag and geographic separation between the bribe payment by the briber to the bribe receiver, and delivery of the good to the briber (S. Li and Wu 2010). The literature has however not distinguished between relation centrism in family and friends as regards the perception of corruption as a business obstacle. When investigating the spheres of trust and the role of relationships in the community and workplace, the deepest bonds are usually for the family (Gonzalez et al. 2018), followed by friendship bonds (Jamieson et al. 2006), and the weakest bonds are usually generalized trust in other members of society (Bjørnskov 2007; Tan and Tambyah 2011).

Generalized trust beliefs likely facilitate the establishment of informal cooperation and the forging of alliances (Ellwardt et al. 2012). Institutional theory posits that a firm's corruption environment encompasses two dimensions – the formal corruption environment (FCE) and the informal corruption environment (ICE) (Kouznetsov et al. 2019). In many cases, corruption thrives within the context of the informal institutions in developing and/or low- and middle-income countries (Mateev et al. 2024). Informal networks and relationships based on kinship, ethnicity, or personal connections may influence access to resources, services, and opportunities. Nepotism, favoritism, and bribery are often facilitated through these informal channels, allowing individuals to circumvent official procedures and regulations for personal gain. As a result, corruption becomes entrenched within the fabric of society, eroding trust in public institutions, undermining the rule of law, and hindering economic progress. The association between corruption and informal institutions can be particularly pronounced in developing and low- to middle-income countries, where formal governance structures may be weak or ineffective (Mateev et al. 2024). In the absence of robust legal and regulatory frameworks, informal networks often serve as alternative mechanisms for resolving disputes, allocating resources, and exercising authority. However, these informal systems may lack transparency, accountability, and safeguards against abuse, creating opportunities for corruption to flourish unchecked.

The social network is important to this enquiry because research from Korea shows that family control of firms may facilitate corruption (Oh et al. 2019), and research from India suggests that firms with closer social networks with the government are more likely to engage in corruption (Collins et al. 2009). Similarly, in China, the process of building up guanxi (personal networks) for financing business operations has led to the normalization of corruption (Kang et al. 2023; Zhu et al. 2022). For this reason, we introduce a new construct “relation centrism”, defined as the importance that people give to family and friends in society. While generalized trust refers to trust in out-group members of society, relation centrism builds on the concept of in-group centrism to refer to trust in family members or friends (Kruglanski et al. 2006). Such ties and in-group favoritism affect firm activity (Yeung 1997; Zhu et al. 2022) and the literature has defined particularized relational trust as trust found in close social proximity and extended toward people the individual knows from everyday interactions (e.g., family members, friends, neighbors and co-workers) (Zheng et al. 2023). When investigating the spheres of trust and the role of relationships in the community and workplace, the deepest in-group bonds are usually for the family (Gonzalez et al. 2018), followed by friendship bonds (Jamieson et al. 2006), and then out-group bonds for generalized trust in other members of society (Freitag and Traunmüller 2009). We express this relationship in Fig. 1.

Full size image

When people put more importance on family members, they interact more frequently with them in activities and may make decisions in their interests although such decisions are sub-optimal from the broader society (Cruz et al. 2012; Zhu et al. 2022). This type of bond is especially strong in cases of vulnerability and also where there are ties of kinship and affection (Rose-Ackerman 2001; Zhu et al. 2022). A different bond could place importance on friends and can be “interest-based” reflecting shared values and goals (Braithwaite and Levi 1998; Cook 2001). Family and friendship ties, distinct from each other and from the weak ties implied by generalized trust, often hold greater strength and may represent in-group connections grounded in trust and relationships, albeit with potential variations in their individual effects. Similarly, in societies valuing relation-centric values, firms might perceive corruption as a more significant obstacle due to the robust emotional or interest-based ties within family or friends. This could potentially isolate individuals without strong family or friend ties, making them feel that corruption is a greater business obstacle.

Banfield (1967) argued that economic performance in southern Italy was depressed because of an inability to trust economic exchange partners outside a close family network. He called it “amoral familism” and showed that trust among members of a business group can functionally exist and be promoted by mistrust of other societal groups. Thus, “bonding” or “strong ties” social capital might lead to strong trust within groups but not between them; indeed, to the extent that strong ties are exclusive, they may promote disintegration and distrust between groups which could lead to increased corruption (Harris 2007). For these reasons, we anticipate a negative correlation between generalized out-group trust and in-group family centrism. Furthermore, we aim to explore whether heightened levels of generalized trust within society, as well as family or friend centrism, correspond to increased perceptions of corruption as a business obstacle for firms. These bonds range from generalized trust in society members to deeper connections within friend groups and the deepest affiliations with family members, forming the basis for our initial hypotheses:

Getz and Volkema (2001) use Hofstede’s cultural dimensions to show that culture can allow corruption to be tolerated in some regions (Hofstede 2011). For example, the dimensions of uncertainty avoidance moderated the relationship between economic adversity and corruption in some regions, whereas power distance and uncertainty avoidance were positively associated with corruption in other regions. This implies that the relationships we have hypothesized are likely to vary across different regions and countries. Regional cultures vary in the degree to which people – individually and within their organizations – trust and interact with one another, which is why regional outcomes vary (Malecki 2012). Generalized trust has been found to have a relationship with ethnic nepotism in Africa (Zerfu et al. 2009), and political participation in Asia (Kim 2014). As broad categorizations, we group the countries in Africa and Asia separately to exploratorily examine the relationships of the role of generalized trust and relation centrism for the extent to which firms perceive corruption as a business obstacle in these two continents given their geographical and cultural differences (Dunford and Liu 2017; Gohou and Soumaré, 2012; Lipshitz and Raveh 1998).

Corruption prevails in numerous lower- and middle-income countries due to underdeveloped political landscapes that foster a culture conducive to corruption. Well-connected firms often leverage these conditions for personal gain (Beesley and Hawkins 2022; Petrou and Thanos 2014; Taylor et al. 2022; Wang et al. 2018). Additionally, cultural, religious, and contextual disparities in these nations significantly influence perceived corruption, impacting how firms view it as a business obstacle (Adomako et al. 2021; Budak and Rajh 2014; Mensah 2014). Given these regional disparities, differences in relation centrism and generalized trust play varying roles in shaping corruption as a business hurdle. Focusing on lower- and middle-income countries, we categorize these nations into four regions based on the United Nations geoscheme,¹ aligning with geographical, political, economic, and historical cultural contexts: South and Central Asia, Middle East and North Africa (MENA), Sub-Saharan Africa, and ASEAN (WorldAtlas, 2023). While these regions share commonalities, disparities in cultural norms, per-capita incomes, and legal institutions are prevalent. These disparities should facilitate different in-group and out-group dynamics due to different cultural bases for a wider range of market-based transactions thus significantly impacting how generalized trust, relation centrism in family or friends, and firms' perceptions of corruption as a business obstacle are interconnected, forming the basis for our ensuing hypotheses:

Governance generally refers to “rule by the rulers” under some defined laws, processes and vivid authority and good governance, more specifically, entails effectiveness and efficiency in state administration (Kaufmann et al. 2011; World Bank 2017). Therefore, while good governance tends to imply impartiality and effectiveness in government, poor governance creates more incentives and chances for corruption. The role of governance in determining corruption is well documented. Kaufmann et al. (1999) contended that the predominant cause of corruption is weak governance. Meagher et al. (2005) made similar conclusions from a study in Bulgaria where they found that corruption was due to failures in regulatory quality and accountability in designing policies. Analogously, Shim and Eom (2008) and Dreher et al. (2009) also showed that good governance leads to a decrease in corruption. Likewise, Attila (2011) investigated the relationship between corruption and regulation, bureaucracy and political structures and concluded that better public institutions are connected to lower levels of corruption. Moreover, Goel et al. (2012) explored the impact of economic freedom, bureaucratic quality, democratic accountability, and law and order on corruption and found an inverse relationship between corruption and institutional factors.

The quality of governance is important for this research because related studies reveal that governance, economic, and socio-political features considerably affect anti-corruption disclosure (Manes-Rossi et al. 2023), and that managers should avoid obfuscating governance records to ensure greater accountability (Ferri et al. 2023). In the same vein, Bjørnskov (2007), who uses social confidence as a measure of institutional quality, posited that countries with a high level of social trust and confidence are more likely to address corruption problems effectively. Therefore, the level of social trust in a society and its relation to the perception of corruption as a business obstacle is influenced by the quality of governance that exists. Firms expect a country with high-quality governance to have less corruption as individuals will have faith in institutions and thus be less reliant on relations to get things done. Rothstein (2011), argues that government institutions, especially courts and law-enforcement offices matter in explaining why generalized or particularized trust is high in society.

Motivated by theory, this paper focuses on three aspects of governance: legal institutional quality, political stability, and regularity quality as corruption is a social, political and economic construct that is related to a country’s legal, political and social systems (Kaufmann et al. 1999; Keefer and Knack 1997). High quality legal institutions imply that the public has confidence that everyone will be treated equally under the law (Berkel et al. 2022; Lv et al. 2021). High political stability means that there is less politically motivated violence and terrorism within a country (Khurana et al. 2022; Miao et al. 2022), while high regularity quality means that the rules formulated by the state promote private sector development (Boudreaux et al. 2022; Kaufmann et al. 1999; Treisman 2000). Although these country-level aspects of governance are important for all economies, they are crucial for the less developed and emerging countries that we sample (Keefer and Knack 1997; Treisman 2000). We, therefore, propose that better governance indicators will reduce the perception of corruption as a business obstacle through generalized trust and relation centrism. Consequently, we make the following hypotheses:

A comprehensive list and descriptive statistics of all our variables is presented in Table 1. In the sample, the firm-level data on corruption perception, firm characteristics, and their views on legal institutional quality and business environment are from the 2013–2016 World Bank Enterprise Survey (WBES) database of the World Bank Group.² The sample is restricted to firms from 20 lower- and middle-income countries where data was collected using the global methodology³ from 2013–2016 to ensure data uniformity. The Enterprise Surveys are administered to a representative sample of firms in the non-agricultural formal private economy and are firm-level data. The final sample consists of 16,785 firms from the manufacturing, services, transportation and construction sectors. Public utilities, government services, health care, and financial services sectors are not included in the sample. The WBES is based on a wide array of qualitative and quantitative information through face-to-face interviews with firm managers and owners regarding the business environment in their countries and the productivity of their firms. The topics covered in the WBES include corruption, infrastructure, trade, finance, regulations, taxes and business licensing, crime and informality, finance, innovation, labor, and perceptions about obstacles to doing business (World Bank 2017).

In addition to the WBES database, we make use of the World Value Survey (WVS) for the corresponding years to capture generalized trust and family and friend centrism at the national level. The WVS consists of nationally representative surveys conducted in countries that contain almost 90% of the world’s population, using a common questionnaire. The WVS is the largest non-commercial, cross-national, time series investigation of human beliefs and values. Variables in the WVS database explore the beliefs, values and motivations of people throughout the world (Alemn and Woods 2016; Harris 2007; Rose‐Ackerman 2001). Finally, we gathered two indicators of country governance, political stability and regulatory quality, from the World Governance Indicators (WGI). The WGI is a research dataset summarizing the views on the quality of governance provided by a large number of enterprise, citizen and expert survey respondents in industrial and developing countries (Kaufmann et al. 2011). The data are gathered from several survey institutes, think tanks, non-governmental organizations, international organizations, and private sector firms and scaled from -2.5 to 2.5. For these indicators, higher scores reflect higher-quality country governance, and lower scores reflect lower-quality country governance. Table 1 provides more details regarding each variable’s definition.

This study posits that higher levels of generalized trust in society and higher levels of relation centrism are associated with higher levels of the extent to which firms perceive corruption as a business obstacle. It also posits that there are regional differences in this regard, and that the quality of governance lessens the effect of generalized trust and relation centrism on the extent to which firms perceive corruption as a business obstacle. Empirically, it is tempting to test this proposition utilizing workhorse multivariate (multiple and multivariate) regression models. However, the outcome variable for the perception of corruption as a business obstacle used in this study is measured on a Likert-type ordinal scale which is multinomial distributed (see 3.2.1), where the natural evolution of the data drives the unobserved (latent) process of the outcome variable as it progressively moves towards higher thresholds of the obstacle corruption places on the operations of business organizations. For this reason, we employed the Least Absolute Shrinkage and Selection Operator (LASSO) model reduction method to help select the most important variables that are truly essential in influencing the Likert-type ordinal scale variable of corruption as a business obstacle (we have also included an explanation of the LASSO reduction method in the appendix). Consequently, we utilized the LASSO method for the variable selection of the baseline model to determine the nexus between the perception of corruption as a business obstacle, general trust and relation centrism covariates as follows:\({\text{where }}{\mathbf{X^{\prime}}}_{i} {{\varvec{\upbeta}}} = \beta_{1} x_{i1} + \beta_{2} x_{i2} + ... + \beta_{k} x_{ik}\),\(y_{i}^{*}\) is a latent variable ranged \(- \infty\) to \(+ \infty\) on the \(i^{th}\) observation, \({\mathbf{X}}\) denotes \(n \times K\) of explanatory variables, \(\varepsilon_{i}\) is the error term. Assuming that our data consists of \(n\) independent countries facing J-ordered alternative of obstacles to corruption, such that:where \(\alpha_{1} < \alpha_{2} < \alpha_{3} ...\alpha_{J - 1}\). Notice that the obstacles to corruption in country \(i\),\(y_{i}\), is observed in one of the \(J\) categories demarcated by the cutoff points. Consequently, for a model with m-alternative ordered alternatives, that alternative \(j\) for corruption is observed in country \(i\) is expressed as follows:where F denotes the communitive density function of \(\varepsilon_{i}\). The coefficients can be identified by both ordered logit and ordered probit models. However, the error term, \(\varepsilon\), follows a logistic distribution for the logit model with \(F(z) = e^{z} /(1 - e^{z} );\) while it takes the form of a standard normal distribution with \(F(.) = \Phi (.)\) for the probit model.⁴ The empirical estimation of Eq. (1) render estimators of \(\beta\) whose sign can determine whether the unobserved variable, \(y_{i}^{*}\), increases or decreases. Consequently, the marginal effect of the \(j^{th}\) covariate, from Eqs. (1) and (3), is:

Thus, the estimation strategy of this paper including the modelling framework and model specification is underpinned by theoretical and empirical literature and aided by machine learning via LASSO model reduction (Belloni et al. 2012; Tibshirani 1996). We estimate the equations utilising the Stata software, we employed the ‘robust’ command to ensure that the standard errors were unbiased and to address the problem of heteroscedasticity.

We have expanded the connections within the trust and corruption literature (Hatak et al. 2015; Keig et al. 2015; Raz et al. 2023) by emphasizing the significance attributed to trust in family, friends, or the broader community (Bullough et al. 2017; Harris 2007; Uslaner 2004). There are three possible interpretations of our results. The first is that medium ties do not create pressures to engage in corrupt behavior because they do not imply significant obligations towards members of the family in-group, nor do they imply substantial trust in out-group members. The second, and in our view more likely interpretation, is that the moderate ties implied by friend centrism serve to deter severe forms of corruption that could pose obstacles to firms in such contexts. This interpretation aligns with consistently lower coefficients for corruption as a business obstacle in Tables 5 and 6, columns 4 and 5. The third interpretation, related to the second, may stem from not having to rely on dense kinship ties to perform business tasks. Markedly, the results for family centrism in Tables 7 and 8 highlight that family centrism generally outweighs friend centrism in firms' perceptions of corruption obstacles.

Our focus has revealed that higher out-group generalized trust, and to a lesser extent, in-group family centrism, are most frequently associated with greater perceptions of corruption hindering business operations. We have also discovered that moderate relation centrism, particularly friend centrism, is most frequently associated with a decrease in the extent to which firms perceive corruption as an obstacle in their operations. Our contribution does not end there, however, as we have also found that regional disparities exist in the nature of these relationships, but that family centrism generally outweighs friend centrism in firms' perceptions of corruption obstacles. Regionally, family centrism is associated with severe perceptions of corruption as a business obstacle in the African sample but not in the Asian sample. Thus, two significant contributions of this paper are that: (i) friend-centric approaches, particularly in organizational or group settings, are more likely than generalized trust and family-centric approaches to maintain checks and balances in lower- and middle-income countries, ensuring that firm relationships and decisions remain ethical, unbiased, and effective; and (ii) family-centric approaches, which emphasize ordered relationships, respect for the rule of law, and adherence to social norms and roles, are more likely to have a negative relationship with severe perceptions of corruption as a business obstacle. Furthermore, we have contributed to the anti-corruption literature by demonstrating the crucial role of political stability in moderating the impact of trust spheres and relationship dynamics on the perception of corruption as a business impediment across regions (Hauser 2019; Jong and Ees 2014).

The study's limitation lies in our examination of lower- and middle-income countries, where strong family ties often play a crucial role in surmounting societal and structural challenges (Mertzanis 2019). Additionally, we remain uncertain about the interplay: whether political stability diminishes the necessity for corruption through generalized trust and relation centrism, or if, in stable political contexts, these elements are employed to reduce corruption. Nevertheless, the robust findings we've presented hold significance, highlighting the crucial need for firms to acknowledge and tackle in-group and out-group dynamics. This awareness is essential not only to foster positive group dynamics but also to alleviate the adverse impacts of exclusion and discrimination within their operational spheres.

Consequently, in societies where family and friend centrism, which focus on relationships, demonstrate a positive and significant relationship with corruption as a business obstacle, firms must ensure that their employees make unbiased decisions devoid of influence from family or other in-group considerations. These findings hold particular importance for firms situated in lower- and middle-income countries, where ethnic and group ties profoundly impact business activities. In such contexts, relationship-centric interactions occur frequently and may significantly impact corruption as a business obstacle and the strategies adopted to mitigate it. Future research could benefit from utilizing longer time-series data across a wider array of countries, as this may reveal deeper dynamics in the evolving nature of regional disparities in how generalized trust and relation centrism impact firms' perceptions of corruption obstacles over time.