THE INSTITUTIONAL LOGICS OF CORRUPTION: AN ANALYSIS BY THE DISCURSIVE PATTERNS OF MANAGERS IN BRAZIL

Our aim was to verify the ‘institutional logics’ of corruption through the discursive patterns of managers. We conducted an exploratory survey using a qualitative–quantitative method. The interviews were submitted to content analysis, using the Atlas.TI software, and the questionnaire was applied by SurveyMonkey. As a methodological contribution, four categories and levels for analyzing the logic of corruption were constructed: Individual, Organization, Networks, and Interinstitutional System. We infer that by understanding the institutional logic of corruption this can promote a transformation of corrupt behavior and an institutional change by individual cognition that can lead to practical change in society.


Introduction
Since the beginning of this century, news about corruption in Brazil has become increasingly recurrent, marking the country's political and business landscape. The term corruption is treated in various approaches in different fields of knowledge. Silva (1999) highlights its polysemic character, highlighting the lack of a functional definition. The study of corruption in public organizations takes place with an approach that is causally related to state activities and especially to its monopolistic and discretionary power. Some aspects of government activities provide a broad scope for corruption, including regulations and authorizations, levying fees, spending decisions, forecasting goods and services below-market prices, and political parties financing (Tanzi, 1998).
Corruption-related problems, regardless of whether the sphere is public or private, tend to be linked to human actions and can be analyzed from a variety of points of view, which promotes a variety of understandings of corruption acceptance levels, as graded smaller or larger in certain areas (Torsello & Vernard, 2015). Corruption is still perceived in different ways by individuals, involving deep individual subjectivities and intense feelings (Brei, 1996). Many academic works in the national literature (Borini & Grisi, 2009;Santos, Guevara, & Amorim, 2013;Melo, Sampaio, & Oliveira, 2015) and international (Belousova, Goel, & Korhonen, 2014;Cooray & Schneider, 2015;Lio & Lee, 2016) start from a quantitative perspective to analyze the phenomenon.
In the literature of organizational studies in Brazil, there is a gap regarding studies that involve the interpretation of corruption by managers working in our country. In turn, in developing the perspective of institutional logics, central researchers have proposed that an institutional logic is composed of the following key elements: a cognitive schema, normative expectations, and material practices (Friedland & Alford, 1991;Thornton & Ocasio, 2008). Based on these considerations, we formulated the following question that guides this research: How can the interpretations of public and private managers of the term corruption lead us to understand the institutional logic of corruption in Brazil?
Following Doty and Glick (1994), we constructed our study's typology using these criteria: (1) identification of constructs, (2) specification of the relationships between the constructs, and (3) how these relationships are falsifiable. In conjunction with Jancsics' model (2014) and with the aid of the qualitative analysis software Atlas.TI, we applied the concept of typologies to analyze the discursive pattern of 20 public health and education managers, considered the most relevant areas in the public budget, residing in the city of Uberlândia. It is a city of about 700,000 residents and is located in the southwest of Minas Gerais state that ranks as the second most populous state of Brazil, the third by gross domestic product (GDP), and the fourth largest by area in the country.
To deepen the question and understand if the logic of corruption is the same in different geographic localities and sectors of Brazil, a questionnaire was applied to 82 private managers in the city of Curitiba, Paraná. Despite being one of the smallest states in the country, Paraná has the fifth largest Gross Domestic Product (GDP) and the third-largest added value of the Manufacturing Industry. The Curitiba Metropolitan area comprises 29 municipalities with a total population of over 3.6 million (IBGE, 2020), making it the nineth most populous metropolitan area in the country. In turn, the SurveyMonkey online survey tool was utilized in this phase and provided direct download of collected data, facilitating multivariate analysis of applied questionnaires.
The article is structured in five sections. After this introduction, we present the theoretical and conceptual perspectives of corruption. We then describe the methodological procedures adopted in this research, and in the fourth section, we discuss the results of the research. We conclude by presenting our final remarks.

Theoretical Perspectives of Corruption
The theoretical perspectives and concepts of corruption vary according to the paradigms that support them. Backed by a realistic paradigm, the definitions, such as put forward by Transparency International (2009), The United Nations Global Program Against Corruption (2001), and The World Bank Independent Evaluation Group (2006), that corruption is the use from public resources for private gain, represent a classic definition for the term. However, these are totally contrary to the perspective presented by the disciplines that are based on a paradigm of nominalism, which considers the relationships and interactions that can lead to the phenomenon. This conceptualization, which understands corruption as an abuse of public office for private gain (Mashali, 2012) and is realistic, does not reflect a consensus, as Brei (1996) states, especially if we consider the pre-paradigm stage of scientific research. It also a fact that much of the academic approach addresses the negative effects of corruption and is concerned with analyzing the causes and reducing the problem.
Economics takes an approach to the corruption that does not focus on moral issues (Rose-Ackerman, 2006). Specifically, accounting addresses the issue from a legalistic perspective, as administrative facts that are verifiable; that is, actions are recorded by accounting departments to check their legality, legitimacy, and veracity, thus, this system becomes an initial filter in the fight against corruption (Oliveira, 2006). Studies in the areas of sociology and anthropology are supported by a nominalist paradigm interested in moral issues and relationships and interactions, moving away from the conceptual models presented by sciences such as accounting (Alatas, 1986). In political studies, two research agendas stand out: (1) the theory of modernization and that of (2) rational choice and the new institutionalism, which applies in sociology to the study of institutions (Filgueiras, 2009).
Political sciences also address the particularities of the political systems in force in each nation. Considering the diversity of countries and the disparities between democratic and dictatorial regimes, it can be noted that because the latter exercise strict control over corruption, they ultimately direct the level and scope of corruption in an autocratically restricted manner. This explains the relatively low level of corruption in some strong and authoritarian states. This is because these regimes are considered legitimate in the eyes of their citizens, for the control of law and order, for social transformations, and for economic growth. In counterpoint to the analysis of authoritarian regimes, democracy itself may contain traces of incentives for corruption. This is due to the regulation of election campaigns, contention for office, profits from privatization, and many other possible forms of corruption (Amundsen, 1999).
When standards and practices serve the interests of an organization to make a profit at any price, the internal restraints against corruption may break down, Ashforth et al. point out (2008). The occurrence of corruption in a business environment is supported by the inability to mobilize capacity and resources for activities that generate greater value than corruption: in the inadequate selection of labor and in failing to value the ethical and moral formation of the individual to the detriment of qualifications and technique and in the institutionalization of a corporate culture lacking norms that curb illegal practices (Borini & Grisi, 2009).
When researchers study corruption in the organizational context, conceptual problems can be expected to arise, which is due to the wide field of analysis to which the phenomenon is subject (Santos, Guevara, & Amorim, 2013). Several angles for the analysis of organizational studies are possible, whether in the individual, organizational, industrial, national, or international field (Karmann et al., 2016). Thus, conceptual variations will occur according to the research context and the field of knowledge (Figure 1). From the analysis of what has been presented for each field of knowledge, the evasiveness of the conceptualization of the term corruption is easily confirmed. Each area makes its contribution, but in isolation, using inherent models and perspectives. Jancsics (2014) classified studies on corruption from different disciplines into three categories (Figure2).  Corruption is a problem within the organization: "bad apple"

Figure 1 -Summary of theoretical explanatory tables of corruption
Corruption is a systematic product of selection processes: "bad barrel" Corruption is an informal network that acts in formal structures. Source: Jancsics (2014) Corruption analysis levels may vary between macro and micro, where the questions may or may not be similar, but the expected answers will certainly not always be equivalent, due to the particularities of the theme (Connelly & Ones, 2008). This gradation of dimensions is also known as "minor corruption" (e.g., bribery) and "major corruption" (e.g., overpricing of public works). Mashali (2012) states that for "small corruption" or disorganized corruption to occur, a buyer and a seller must exist. For the author, the difference between small and large corruption is that, in the latter case, there is no dependence on individuals for it to happen, because, at this level, it is institutional, that is, it is already inserted in the system.
Cases of major corruption, organized corruption, or aggravated corruption (Heidenheimer, 2002) develop in the confidentiality of negotiations, and would not be known to the general public, although these cases are eventually reported by the mainstream media. The character of this type of corruption is based on the confidential action of managers who seek to conceal rather than address a problem so that it can be properly addressed (Mashali, 2012).
Regarding corruption in the public sector, Mashali (2012) states that it occurs between organizations and public persons or between one or more public sector actors within the private sector. In the public sector, the powers and facilities belong to everyone, so when there is corruption, the burden is on the whole of society. Thus, the involved agent considers his or her position as business and will seek to maximize his or her income (Brei, 1996.). Generally, corruption in this sector is also political, as individuals have the skill, experience, and ability to do their work, but still practice corruption of their own accord (Mashali, 2012).
The possibilities for corruption in the public sector are, for the most part, connected with state activities and especially with its monopolistic and discretionary powers. Regulations and authorizations, fee collection, spending decisions, forecasting of goods and services below-market prices, as well as political parties financing, are some of the aspects of government activities that provide an extensive field for corrupt government action (Tanzi, 1998).
The concept of corruption in the private sector is close to that of "petty corruption" since the abuse of power for personal gain is illegitimate in this sector because, without the involvement of the public sector, it has no direct consequences for society as a whole. Generally, losses do not extend outside the scope of the organization. In these situations, corruption can also be considered technical, as individuals who practice it lack information, skills, and experience (Mashali, 2012). Zyglidopoulos (2016), examining the concepts of corruption that deal with abuse of power, whether in a managerial or political context, designed a proposal for two types of corruption: a first-order model linked to the classic concept of corruption, in which an individual or group abuses a system of rules for private gain; and a second-order model, in which individuals or groups alter rules and norms for their own benefit, similar to the "great corruption" identified by other authors (Tanzi, 1998;Mashali, 2012). Ashfort et al. (2008) maintain that, in the analysis of corruption within organizations, the following levels exist: macro (organizations influence the actions of their members, with organizational cultures operating norms that guide the practices of their members) and micro (corruption is limited to individual and organizational forces that encourage individual corrupt behavior, refraining from factors that demonstrate how organizations become corrupt over time). The authors argue that a macro view contributes to a better understanding of the causes and transformations of the phenomenon.
According to Pinto, Leana, and Pil (2008), corruption in organizations has two dimensions: the first considers the organization to be the direct financial beneficiary of acts of corruption, even if its practitioners acquire indirect benefits, whether financial or personal. The second is the individual dimension, which is related to collusion among the members of an organization, although much of the literature focuses on individual actions at the expense of those collusions.
From these dimensions, organizations can then be classified as (i) an organization of corrupt individuals (OCI) or (ii) a corrupt organization (CO). OCIs emerge from the individual level and are enhanced by interactions between individuals on an organizational level, while COs arise from a coalition of dominant members with actions that primarily benefit the organization (Pinto, Leana, & Pill, 2008).
In addition to manifestations in organizations, corruption is closely related to morality (Filgueiras, 2006), which, in turn, involves judgments that allow the acceptance or rejection of an action by the individual. With this understanding, Filgueiras (2006) distinguishes the manifestation of corruption as between two distinct types of moral judgments: that of value and necessity. In the value judgment, the author encompasses the manifestations of corruption in the political and cultural forms, while as judgments of necessity, the social and economic forms of manifestation of the phenomenon are categorized. From a political perspective, cultural, social, and economic elements must also be considered. The author, based on moral judgments, presents variations in understanding the phenomenon according to contemporary political orders.
In this section, we have presented the different perspectives on corruption from various disciplines, which points to the fact that interdisciplinarity in the study of corruption can contribute to a deeper analysis of the phenomenon and to the understanding of the institutional logics of corruption. There is a richness and diversity in the results found by the different areas of study that contribute substantially to the knowledge on the subject. However, the risk of excessive ambiguity challenges the development of studies that provide a dialogue between disciplines for a broader understanding of the subject. In the next section, we present a methodological typology we created to be tested based on the institutional logic of corruption identified in Brazil and that can be like other countries.

Methodological Procedures
We adopted the qualitative approach in this research at the first stage because it is suitable for studies of social relations, as it gives them a new sensitivity to the empirical study of several issues that relate to the pluralization of life spheres (Flick, 2009). The technique chosen for composition of the research corpus was the semi-structured interview because it allows an increase in the degree of comparison of the statements, making them more structured, as well as with some flexibility to include new questions during the interview (Flick, 2009). In the second stage, we were interested in the differences between geographical regions in Brazil, to see if managers from another state and private sector were having the same view of corruption, so we decided to apply a quantitative study to complement the qualitative result.
Thus, based on the results of the study by Jancsics (2014), which classified studies on corruption from different disciplines, we adopted the ideal type approach to analyze discursive patterns through interviews with 20 public managers in the city of Uberlândia, Minas Gerais and administered a questionnaire to 82 private managers working in Curitiba, Paraná. Following in the subsections we explain the data collection and how the analysis and use of the typology took place.

Data collection and analysis
We defined as research corpus the interviews of public managers working in the areas of health and education, which are two areas that suffer a large negative impact due to corruption cases, from the three spheres of public administration (municipal, state, and federal) of public servants in the city of Uberlândia-MG. The choice of conducting the research in the city of Uberlândia-MG is justified by the city is in the state of Minas Gerais which, according to the Corruption Map (MPF, 2017), is the third Brazilian state with the highest number of cases of corruption. Uberlândia is a city of about 700,000 residents and is located in the southwest of Minas Gerais state that also ranks as the second most populous state of Brazil, the third by gross domestic product (GDP), and the fourth largest by area in the country.
The survey was conducted with a total of twenty civil servants, ten acting at the state level and ten at the federal level, in agencies located in the city of Uberlândia-MG. The interviewed managers work at the Regional Health Superintendence of Minas Gerais -Uberlândia Unit, State Superintendent of Education of Minas Gerais -Uberlândia Unit, Clinical Hospital of the Federal University of Uberlândia and Federal University of Uberlândia.
In addition to the state and federal agencies, the request was also made to conduct research in the municipal organs, since the initial proposal of the research was to be carried out at the three levels of public power in the city of Uberlândia. However, the Municipal Secretary of Education did not authorize the research with the municipal servants for not agreeing with the questions that are part of the questionnaire; and the Municipal Health Department did not respond to our requests, even after three months of constant attempts to receive authorization to develop the research.
The selection of respondents was by convenience, and the number of interviews was determined based on the saturation of information collected in the field (Fontanella et al., 2011, p. 390). For these authors, there is "theoretical saturation when the interaction between research field and the researcher no longer provides elements to guide or deepen theorizing." The interviews were transcribed, totaling 189 pages, and analyzed using coding content analysis techniques recommended by Bardin (2011) and using Atlas.TI support software.
After the exploratory reading of the content, the interviews were classified according to the codes and citations corresponding to the analysis categories, which served to construct the questionnaire, which we decided to apply in another geographical area in Brazil and with private managers to deepen the question and understand if the logic of corruption is the same in different geographic localities and sectors of Brazil. In turn, the SurveyMonkey online survey tool provided direct download of the collected data, facilitating multivariate analysis of the applied questionnaires. The data obtained through the application of the questionnaires were treated using multivariate exploratory factor analysis and cluster analysis techniques and the statistical software SPSS, in version 20. The level of significance adopted was 5%, although it was used only as confirmatory for triangulation and to reinforce the findings of the interviews, being part of a larger project of one of the authors in the city of Curitiba, Paraná.
In Brazil, the state of Paraná is the fifth in the corruption ranking, according to the Corruption Map (MPF, 2017). Despite being one of the smallest states in the country, Paraná has the fifth largest Gross Domestic Product (GDP) and the third-largest added value of the Manufacturing Industry. The Curitiba Metropolitan area comprises 29 municipalities with a total population of over 3.6 million (IBGE, 2020), making it the nineth most populous metropolitan area in the country.

Typologies as methodological framework
The development of the ideal type method by Weber (1946) to explain the structures of traditional, charismatic, and legal authority inaugurated in sociology the use of typologies as a methodology for analyzing social phenomena. In the area of organizational studies, typologies are useful because they allow organizational complexity to be described concisely, or for collective thoughts to be analyzed under the group category (Doty & Glick, 1994). Thus, typologies are viewed as interesting tools for understanding organizational structures and strategies, such as the typologies developed by Miles and Snow in 1978, Mintzberg in 1979and 1983, and Porter in 1980and 1985(Doty & Glick, 1994Thornton, 2004).
According to Doty and Glick (1994, p. 243), typologies are "complex theoretical statements developed to predict variance in dependent variables." However, the authors argue that a problem that generates misinterpretation of the concept of typology is the confusion related to the concepts of a scheme and taxonomy. The latter are considered by the authors to be classification systems responsible for categorizing a phenomenon with mutually exclusive dimensions, based on delimited decision rules. In turn, typology conceptually refers to an interrelated set of ideal types. To clarify the differentiation, they argue that, unlike classification systems, typologies "do not provide decision rules for classifying organizations. Instead, typologies identify multiple ideal types, each representing a single combination of attributes" (Doty & Glick, 1994, p. 232).
For Thornton, Jones, and Kury (2005), the development of ideal types is a method for multidimensional classification of the phenomenon that is not restricted by the events of the selected cases or the characteristics of the selected sample. This feature is extremely relevant to social studies because it enables the use of the ideal type methodology as a guide for the analysis of different empirical cases (Ragin, 1987). Thus, they argue that as a theoretical model, ideal types offer hypotheses that can be tested by examining similarities and discrepancies between ideal types and dependent and independent variables, according to each case under study. In addition, Doty and Glick (1994) accept that typologies may be theories, but for this they must contain the minimal definitions of a theory. For the authors, three criteria must be considered in the theoretical construction: (1) the constructs must be identified, (2) the relationships between the constructs must be specified, and (3) these relationships must be falsifiable.
In the construct's identification process, the authors posit that, starting from the characterization of the ideal types, the dimensions used to describe each ideal type in the theory should be identified. In addition, ideal types must be described in multiple dimensions. Referring to the Weberian conception, Doty and Glick (1994) also point out that, because they are theoretical abstractions, ideal types are not hypotheses, but offer a guide for the construction of hypotheses. Thus, ideal types are considered to be useful for identifying organizations that may exist rather than existing organizations, so they should not be considered as literal descriptions of concrete organizations, but references so that existing organizations can be viewed as more or less similar to what they are.
Regarding the second criterion, the specification of relationships between constructs, the authors argue that typologies are responsible for establishing hypotheses about the similarity of an organization and an ideal type, as well as about their respective dependent variables. For Doty and Glick (1994), typological theories, unlike more traditional ones, do not highlight the hypothetical relationships between first-order constructs and dependent variables, but rather the internal consistency between the first order of constructs with an ideal type, which explains why this internal consistency results in a specific level of the dependent variable(s). As for the third criterion, this is related to the fact that predictions must be subject to fallibility, that is, predictions associated with a typology must be "testable and subject to disconfirmation" (Doty & Glick, 1994).
In the view of Thornton and Ocasio (2008), the use of typologies has several advantages over other methods. This is mainly because ideal types do not represent only the empirical cases of the sample; instead, a typology is not restricted to the events of the selected cases or characteristics. In contrast, the use of typologies makes it possible to specify multiple construct patterns and nonlinear relationships that determine the dependent variable and can be useful for testing theorizations using historical event models. Therefore, the authors consider the typological scheme to be suitable for appreciating cultural phenomena and for providing a guide for developing hypotheses about the effects of institutional change.  Corruption is a problem within the organization: "bad apple" Corruption is a systematic product of selection processes: "bad barrel" Corruption is an informal network that acts in formal structures Source: Adapted Thornton, Ocasio and Lounsbury (2012, p. 56) and Jancsics (2014).

Results and Discussion
Besides allowing the multidimensional classification of the phenomenon under study, the use of typologies encompasses two questions (1) the description of ideal types and (2) the set of statements that relate ideal types with dependent variables. Therefore, ideal types are intended to provide an abstract model that enables the combination of attributes that determine the variables of interest (Thornton, 2002). We consider that a qualitative and quantitative research is relevant for understanding the institutional logic of corruption (Brei, 1996;Ashforth et. al, 2002;Jancsics, 2014) because it can capture discourses in the same conception of Fairclough (2001), considering the interaction process of the actors, who now share the same definitions of reality (Leite, Sousa, Souza, & Costa, 2019).
Thus, from the analysis of interviews and questionnaires, through the defined categories we consider that the corruption process involves different levels: (i) individual: composed of characteristics related to the identity of the subject and the capabilities of the individuals involved in the management process. and their relationships; (ii) organizational: pertinent to the organizational policies, rules and strategies and management activities that structure the internal and external behavior of the organization; (iii) networks: related to the characteristics of interorganizational arrangements, collaborative network provisions, and institutional rules that structure interaction between actors within a context (Granovetter, 2006;Jancsics, 2014;Gieske, Buuren, & Bekkers, 2016) and ( iv) interinstitutional system: refers to the culture and history of the social context capable of influencing behavior through routine and practices taken for granted by individuals over time. Such factors would act as an interactive system or composed of a system of multiple institutional logics (Scott & Bruce, 1994;Thornton, Ocasio, & Lounsbury, 2012), which would mutually influence social practices and the capacity for institutional change to occur. Below we describe each of the constructs, as well as the possibility of relationship between them and falsification of this proposed relationship.

Individual Level
At the individual level, an individual's personality traits and the social context in which they are immersed can influence their behavior (Thornton, Ocasio, & Lounsbury, 2012). From the perspective of institutional logics, there are distinct logics that influence identity; according to Thornton, Ocasio, and Lounsbury (2012), society is an interinstitutional system composed by the logics of family, community, religion, state, market, profession, and corporation, all of which influence practices through the identity of the subjects. For Friedland, Mohr, Roose, & Gardinali (2014), an institutional logic is an analytic troika of object, practice, and subject linked together through dually ordered systems of articulations.
In this category, the moralist current takes the nominalist approach to examine the relationships that can drive the corruption phenomenon (Alatas, 1986;Filgueiras, 2006, Mashali, 2012Karmann et al., 2016). In the interviewees' discursive patterns, the exchange for benefits is described as bad faith, self-benefit, and advantage, as can be seen in some selected excerpts: Brazil (laughs) . . . Ah, what is the concept of corruption? A corrupt person is a person who acts in bad faith, is a person who thinks about himself, right? I think corruption is more, it's acting on its own, right? So, I think this is corruption is . . . for me, a corrupt person is this . . . (Claudia, in an interview, 2017) I think there is one . . . and then it is a matter of working on the individual problem, but we have a tendency to take advantage of the other . . . (Mario, in an interview, 2017) What motivates corruption is self-benefit, going out with some advantage . . . (Janice, in an interview, 2017) It can also be observed that by discursive standards there is a direct association between the practice of corruption and a selfish attitude related to the corrupt act., which Filgueiras (2009) considered as intrinsic in theory of rational choice as people can take an action based in cost benefit analysis.
No doubt remuneration motivates the individual to become corrupted, no doubt that this is money . . . yeah . . . inside the clinic hospital there, they fight over shifts, then we will think they are fighting by shifts (Galeno, in an interview, 2017). I think it happens, for the same reason as any other organization, there is someone at the end of the chain taking advantage. I'll give you an example here . . . I get there at the general service assistant, he has to wash a bathroom at school, he washes the bathroom floor and forgets the wall, this forgetting is intentional, that he knows he has to wash there, see that there it's dirty that there is a bacterial action the same way and so on, it happens like this in all sectors because individuals want to take advantage (Francisco, in an interview, 2017).
At this level, we deduce from our analysis that corruption is the result of antecedent conditions, cognitive style, skills, personality, knowledge, motivation, and social and contextual influences that give the practicing social actor a sense of impunity, constituted by the legal uncertainty existing in the country. Although some authors consider it to be a legal aspect (Karmann, Mauer, Flatten, & Brettel, 2016), there are still few studies that analyze how the formation of identity in a context of legal insecurity is constituted.
I think the first thing is impunity, right, second thing is to think that once I read a text I don't know whose, to think that corruption is just misuse of resources and big . . . (Delilah, in an interview, 2017) I had an example of impunity here. The father comes out breathless because for some personal reason saying that he as a worker and he is rich and I don't know what else, he will not be punished; he can come in here and kill everyone he can come and beat, because the law in Brazil does not work for him . . . (Saulo, in an interview, 2017).
The motivation for existing corruption, which for Jancsics' (2014) actor-rational approach is related to market logic, we believe is largely derived from the individual's personality as well, being linked to financial issues using words such as "values" and "ethical," which indicate that the phenomenon enters the moral field. However, there are still issues related to the identity of the subject, who needs to be respected or acquire status in front of his or her family, which need further study.
In speaking of the exchange of favors as a form of corruption, one of the interviewees (Matthew) used the term "breaking a branch," which can be associated with the exchange of favors, but also with bonds of friendship or sympathy. Thus, we consider that there is a duality of the identity level of the subject and the social context in which the prevalence of self-interest is reinforced (capitalist market). In such a case, the corrupt act is conceived as a normal act by the absence of a justice that can ensure adequate punishment of offenders (sense of impunity), as well as reinforcement of an individual sense of power in decision-making and action, which in some cases involves emotion in the process and justification for the corrupt act, especially in countries that are in the process of development like Brazil and that are represented by laws that are not enforced by the juridical systems to everyone in society in equal way.
The questionnaire asked a series of questions to determine the motivation for the practice of corruption in the private sector by Brazilian manager's cognition: (1) Was the act of "taking advantage whenever it is in their own interest" to be institutionalized in Brazil? (2) Because the punishment is irrelevant to the amount of benefit gained from wrongful acts? (3) Due to the low occurrence of enforcement? (4) Due to flawed legislation? The results showed that 35 respondents considered the motivation to be because corruption was advantageous and 25 responded because the punishment was irrelevant. The others stated that the motivation was the low occurrence of oversight. That is, 73% of private managers followed the discursive patterns of public managers and the others related to the lack of control and sense of impunity resulting from this.
In this sense, corruption as an individual behavior arising from the abuse of power is similar to the first order corruption model, as presented by Zyglidopoulos (2016). For the author, as for the interviewee, in this model of corruption, abuse of the system and rules for private gains are both commonplace. When considering corruption from the subjective point of view of the individual and his or her cognition, it is advisable to adopt a research agenda that analyzes the related issues of moral, economic, and social consequence, which will be concerned also with the emotions (or absence thereof) involved in the practice of corruption. Such studies will draw upon different disciplines and organizational theory, specifically, to better understand the phenomenon in the routines of organizations. These issues will be discussed below at the organizational level.

Organizational Level
In a capitalist society, where market logic is still dominant and has legitimacy, the organizational context is seen as appropriate. Accordingly, employees realize that the environment in which they work encourages, recognizes, respects, and rewards those who are creative. As a result, employees often act in risky ways to preserve or improve a company's market position and transaction efficiency. For Thornton, Ocasio, and Lounsbury (2012), the source of market identity is "faceless" (p. 73), as an association for which any practice is valid to achieve the primary goal or increase profitability. According to Ashforth et al. (2008), when norms and practices serve the interests of an organization to make profits at any price, such norms may be breached, justifying corruption. We observe in the excerpts below an association with these characteristics: . . . maybe you have a missing employee who doesn't work well and the boss is turning a blind eye, leaving, and that generates in that person, as in the colleagues, a feeling that everyone can do everything, everything is fine, so I'm not going to follow any rules (Joaquim, in interview, 2017). We had a recent case here at the university that culminated in the dismissal of two doctors, public case, that we must talk not about speculation, real public cases, because we know that politics and management generates a lot of speculation around. And we know that the city of Uberlândia is a city that still has some Coronelista traits there, we know that it has a lot of influence on that and then people have a lot of political influence, big businessmen, farmers, of two doctors of that organization that were, got involved in a bidding corruption scheme (Alzira, in interview, 2017). as the leadership acts, I think it mirrors for those below this leadership following this leader in the organization, actions are important maybe because I believe she's creating a mechanism of a corruptible path (Mateus, in interview, 2017).
In the discursive patterns of the interviewees, organizational practices were associated with the corruption and breaking of norms in the first part of a poor business environment or, according to Borini and Grisi (2009), in the inadequate selection of manpower, in failing to value ethical training and individual morality to the detriment of technical qualification, and in the institutionalization of a corporate environment with norms that do not restrain illegal practices.
Organizational corruption is also seen in the discourse pattern of the public manager in the second section who was selected because of policy and interest, and dismissal of employees can be conceived as corruption that was treated as a problem within the organization: that is, the "bad apple" as defined as the ideal type for the organizational level, following Jancsics (2014). However, another interviewee reported the need for administrative proceedings in the agencies and greater difficulty in dismissing a public official than a private company employee in Brazil, what is a legal aspect (Karmann, Mauer, Flatten, & Brettel, 2016).
The following are questions addressed in the questionnaire to understand the organizational practices of private managers: (1) Does the company allow the purchase and sale of goods without an invoice? (2) Is it common for companies not to report cash transactions? (3) Do stories often justify certain unlawful behavior by employees or managers themselves? (4) Do you notice cases of buying and selling diverted goods? For all responses, the private managers confirmed the existence of these practices, and questions 2 and 3 had 16 managers and 9 managers, respectively, who pointed to 100% occurrence in their companies. For deviated purchases, none of the managers recognized 100%, and only one considered 90%. For purchase and sale without invoice, 11 also stated that this is a common practice in their companies, even with the tight integration of the revenue system with banking systems and greater enforcement by electronic means in Brazil in recent years.
The private managers were also asked which area within the organization is most corrupt, and they pointed first to the financial area and second to the purchasing area. Within any company, the ability to establish relationships, the presence of skills, policies, and practices aimed at generating and sustaining internal resources, along with networks that facilitate meaningful connections, are valuable attributes for organizational change and innovation. We also asked the managers to appraise the leader's influence on the organization; in the views of most, a weak or apparently incompetent leader would also be a factor in employee misconduct, which they considered to be equivalent to a lack of controls.
Thus, we suggest that further studies be conducted that relate the leadership profile to fraud and illicit behavior in organizations. While most findings in studies of corruption find that corrupt employees are generally prominent within the organization, it should be noted that an analysis of the correlation between the leader and the consolidation of corruption in the environment and organizational culture in which he or she leads is relevant. Moreover, studies that analyze the permanence of corrupt practices in organizations, even with stricter controls, are also relevant for understanding this phenomenon.

Relational Level
For Jancsics (2014), at the relational level, corruption is a systematic product of selective processes: the "bad barrel" idea. There is a consensus in organizational studies that cooperation networks are responsible for the accumulation of new information, learning, and organizational resources and are also conceived as necessary for their growth, internationalization, and increasing their prestige and market power (Granovetter, 2006;Jancsics, 2014;Gieske, Buuren, & Bekkers, 2016). However, it is apparent that at the relational level corruption increases its degree of complexity, with ties from relatives of varying degrees, friendships, and involvement with politicians to favor the legal environment (Karmann, Mauer, Flatten, & Brettel, 2016). These questions can be observed in discursive patterns: . . . you take diversions of millions all the time and become a network, a very complex web that you can't even know who the mastermind of it is all (Alzira, in an interview, 2017). In these delations there, the delations are all there. Ah I think it's like this when I see it like this, for example, there goes out the story of money in Gedel's apartment, you look and talk like this small fish and if you take the chief then, what's not going find, today I'm more or less like this. Oh and there are people who endorse Sérgio Moro, not behind there is also something, his wife works for I don't know who and I don't know who, who is from PSDB (a Brazilian political party). So, oh, you won't investigate this one because this was Lula's friend at the time of the government, because I don't know what shovel (Paula, in an interview, 2017). These private company practices, I think 99% happen and the government is involved, because only the company alone, it cannot adopt this practice a company with another, private company, they don't have a business relationship, I think there's no way only private organizations acting corruption practice, so almost every time a company is involved in corruption the public government is involved, they are together in a relational (Galeno, in an interview, 2017).
In theory, relational levels are generally characterized by free information exchange and constructive dialogue, and trust is an important asset of networks, as it reduces uncertainty arising from conflicts of interest or opportunistic behavior (Granovetter, 2006;Jancsics, 2014;Gieske, Buuren, & Bekkers, 2016).
The networks in this case may coincide with OCIs that emerge from the individual level and are enhanced by interactions between individuals on an organizational level (Pinto, Leana, & Pill, 2008). However, when associated with corruption practices, networks become powerful tools of social manipulation. According to discursive standards, ties between families that have political power and access to the media also become important links in corruption networks: linked to big political groups, so how, if we are talking, the media that is in the hands of the scammers, the corrupt, how are we going to say that this media has no criticism? Of course, that even then so it is a fraudulent media, is a dirty media (Renato, in an interview, 2017). The media is either biased to one side or the other, the corruption itself also picks up this media issue, maybe you have an organization there, a company that is doing something wrong and it ends up hiring someone to defend or paying someone to stay quiet and do not report it (Elvira, in interview, 2017).
Finally, the questionnaire applied to private managers asked the following questions: (1) Has your company ever offered a bribe to omit illegal practices? (2) Do you often ask about unlawful ways of lowering tax payments on your accounting? The results were discrepant; while 74.39% of private managers (61 respondents) said they question their accountants about illicit ways to lower tax payments, only 20.73% (17 respondents) said they had already offered bribes.
The questions were asked because we assume that one of the objective cost factors for organizations in the institutional environment in Brazil is the tax burden, and we also assume that accounting professionals are usually involved in corruption networks. It is suggested that further studies be conducted to understand the formation and maintenance of networks of illicit practices to better understand the connections that are formed and strengthen the continuity of these practices at the societal level, which will be dealt with below.

Interinstitutional System Level
For Thornton, Ocasio, and Lounsbury (2012), society can be seen as an interinstitutional system composed of six ideal types of logic: family, religion, state, market, profession, and corporation, as already noted. More recent studies have brought the logic of community and sustainability into this system. When we look at corruption from the perspective of the interinstitutional system, we consider that all of these logics influence the process and practice of corruption that makes it institutionalized in some countries more than others. For example, in families where corruption is already seen as normal there are a normal path of perpetuating corruption because the pattern of the father was to get fortune by corruption acts. In such a family heritage, corrupt practices will be seen as the right course of action, although if one of the heirs begins to practice a particular religion that restricts this practice, there will be a conflict in the identity of the subject, and often the latter will not engage in corrupt practices as more confidence he has in his religion logic that this is not an adequate way of act.
In coding, while interest, advantage, and morality appeared 48 times in the interviewees' discourse, culture appeared 74 times as responsible for corruption. Thus, we observe that for managers in Brazil, corruption is a consequence of our culture.
Below are two excerpts from the interviews, but most of them confirm the perception of public managers through their shared discursive pattern: . . . I think what leads to corruption in Brazil is a culture; a Brazilian is kind of corrupt by nature. It's kind of predestined to be corrupt . . . The Brazilian culture, the national culture. I think so, first, as I said before, due to the legislation itself, it is very confusing, it seems that it is already made to be disrespected and to create, creating loopholes and everything and we see today, the Congress, the Senate, working precisely for this, to create legislation to favor the corrupt, so there is a whole culture associated in a formalization of this corruption. Institutionalization is what we have had lately. (José, in an interview, 2017) We have, I think there is a cultural model, a cultural thing, and this culture is rooted in certain segments, certain places, it even gets complicated to think about it. (Debora, in an interview, 2017) If we observe that logics are composed of the following key elements: cognitive schema, normative expectations, and material practices (Friedland & Alford, 1991;Thornton & Ocasio, 2008), we come to understand that as long as the individual in their cognitive schema continues to believe that corruption is already part of our culture, normative expectations are reinforced and materialized through corruption practices accepted as normal because they are part of the Brazilian culture and mechanism of the perpetuation of corruption.
At this level, there is also an association of corruption as a breach of norms supported by a legalistic paradigm (Karmann, Mauer, Flatten, & Brettel, 2016), with the interviewees' discourse associating corruption with rule breaking.
It has much stricter and more complex laws and these simple people don't always pay attention and take it very seriously. I don't know if it's because they think that on campus, they don't have to follow the traffic code, for example, but so generally here in Brazil at least the laws they exist, but often they do not work (Joaquim, in an interview, 2017).
I realize this, as something like that I don't know if you can call it a lack of ethics, because the laws they exist, the rules they exist, to be fulfilled, to be followed in Brazil, but we have there in society is not only within the public power but even within the work itself is this living with breaking rules, with disrespect for the law work (Galeno, in an interview, 2017).
As we have already seen in the networks section, the tax burden is one of the factors that lead private managers to illicit practices. Regarding the question whether high tax rates in Brazil contribute to corruption, private managers in their entirety said yes. Still related to the interinstitutional system, but which we can relate to the other levels, the discursive patterns indicated that mechanisms are created to serve certain groups, but that this generates a social cost, just as does tax evasion.
Look, I am in the group that everyone thinks is corrupt because the other corruption that we understand, so, that is for a better collective, for a greater coexistence that sometimes you create corruption mechanisms for what? To meet a small group (Francisco, in an interview, 2017).
So in the political area I see as a big loss the embezzlement, often money that would go to building a school, a hospital for asphalt reform, something like that, if this is diverted by corruption for other purposes, that ends up harming the population in general (Joaquim, in an interview, 2017).
The damage that I think it brings, right, to society, in general, this lack of resources, we live, if resources were better managed, with responsibility, for example, in the matter of medicines, right? Work begins on a hospital and does not end, right, and then comes deviation of money, right (Francisco, in an interview, 2017).
I think corruption is very much linked to the dishonesty of the person, right? Because the corrupt, he is dishonest, right? And being dishonest, he's hurting someone, so you have no doubt, right? Sometimes you break a rule of law, it will not hurt. Sometimes it can benefit, but the term corruption is sure to hurt someone or hurt many people (Galeno, in an interview, 2017).
As noted in the interview excerpts, the discursive pattern reveals that there is a shared sense of a high social cost generated by maintaining the practice of corruption. But in the last excerpt the enunciator, while linking the concept of corruption to dishonest practice, brings semantics to lexical selection. "Sometimes you break a rule of law, it won't hurt. Sometimes it can benefit" by minimizing the effects of rule breaking. This interviewee's notion that corruption can be beneficial is sometimes in line with the conception of corruption present in the revisionist current, which is present in the perspective of authors such as Bardhan (1997), who use examples such as the "golden age" where there is an association between economic growth and corruption in disobedience to state and local laws.
However, the discourses that understand corruption as a social cost is pointed out by Brei (1996), understand that in the conception of public corruption, the powers and facilities belong to everyone, so when there is the practice of corruption, the onus is on the society. This study, potentially, will contribute to knowledge about the interpretation of corruption by managers about the different understandings of the subject and analysis at the different analytical levels proposed. With the results of this research, besides knowing the interpretations of managers, society will be able to glimpse cultural changes through individual reflections on the phenomenon. Research yet is considered necessary to analyze the influence  ISSN 1982ISSN -2596 of the media on the change of practice and cognition of individuals so that institutional change is possible, as well as measurement and analysis of the social costs involved.

Final Remarks and Research Agenda
This research makes theoretical and methodological contributions and aims to contribute to reflections on changing practices that may make the Brazilian context more transparent. Also, in theoretical field we contribute by the use of ideal type based in corruption studies and organizational institutionalism theory as we considered that the corruption process involves different levels of analysis: (i) individual; (ii) organizational; (iii) networks (Granovetter, 2006;Jancsics, 2014;Gieske, Buuren, & Bekkers, 2016) and (iv) interinstitutional system, but that they are interrelated, and they are also in harmonic way acting in the society.
The results show that private managers' definitions of corruption are not contrary to a dominant discursive of public severs in Brazil order and suit the act indifferent the place geographical and sector analyzed since they show that they are constructed from reproductions of truths that ignore the existence of a critical sense. We infer that by understanding better the institutional logic of corruption this can promote a transformation of corrupt behavior within organizations and an institutional change by individual cognition change that can lead to practical change in Brazilian society.
The origin of corruption, its maintenance and possible perpetuation are worrisome issues that arouse the need to understand the phenomenon and, especially, for the exposure and discussion of issues related to the theme, in hopes that from a better understanding, propositions or changes may arise in the attitude of all who build and integrate organizations and society.
As a research agenda, we proposed analyzes the related issues of moral, economic, and social consequences, which will be concerned also with the emotions (or absence thereof) involved in the practice of corruption. Moreover, further studies that relate the leadership profile to fraud and illicit behavior in organizations and that analyze the permanence of corrupt practices in organizations, even with stricter controls. Also, there are studies to be conducted to understand the formation and maintenance of networks of illicit practices. Finally, research is considered necessary to analyze the influence of the media on the change of practice and cognition of individuals so that institutional change is possible, as well as measurement and analysis of the social costs involved.
This study, potentially, will contribute to the knowledge about the interpretation of corruption by public managers on different understandings of the subject, relating these perceptions with the increase of this practice in the various organizations in Brazil, both public and private initiative. With the results of this research, besides knowing the interpretation of managers, society will be able to glimpse the cultural changes that may contribute to individual reflections on the phenomenon, causing changes in public and private management.
As we can notice that Latin America presents bigger corruption costs than those of developed countries, we also suggest conducting research that fills the gaps found and studies considering public and private sector managers in other countries to make a comparison with the logic of corruption found in Brazil.