Propensity to Corrupt Decisions of Pre-Candidates to the 2020 Brazilian Municipal Elections

Objective: to analyze the propensity for corrupt decision-making by pre-candidates for mayor and councilor, analyzing the 2020 Brazilian municipal elections. eoretical framework: the study is based on the behavioral aspect of corruption, as deviant behavior, and intrinsic and extrinsic motivations to corruption, which are subdivided into a propensity to unethical decision-making, individual self-control, moral Administração Pública e Gestão Social, 2022, vol. 14, núm. 1, Enero-Marzo, ISSN: 2175-5787 PDF generado a partir de XML-JATS4R por Redalyc Proyecto académico sin fines de lucro, desarrollado bajo la iniciativa de acceso abierto identity, guilt, and disengagement moral. is behavioral aspect of corruption is supported in the Agency eory and eory of Public Choice article. Methodology: A quantitative, exploratory, and descriptive study. A cross-sectional survey was used in the operationalization, with the data collected through a structured questionnaire applied to a valid sample of 115 pre-candidates for the 2020 municipal elections. e data analysis used univariate and bivariate statistical techniques. Results: (1) a sample has a low propensity for unethical decision-making, high individual self-control, high moral identity, low guilt, and low propensity for moral disengagement. (2) there is a significant and positive association between the variables of individual self-control with guilt. Moral disengagement and moral identity are shown to be associated with moral disengagement. Moreover, moral disengagement positively with the variable of having a previous elective position or not. Originality: e main contribution of the research is the analysis of the theme on municipal legislators and executors in the version of pre-candidates. eoretical and practical contributions: (1) low representation of women even as pre-candidates; (2) the sample considers the 'other' pre-willing politicians to make unethical decision-making, but not themselves; (3) sample with a high capacity for selfcontrol, especially the part of the sample that did not have a previous elective position; (4) women have a higher degree of guilt than men, that is, they are less disposed to unethical decisions or corruption; (5) it can be said that the sample has a low propensity to deviant behaviors, that is, corruption.

1 Introduction e fight against corruption has been a banner for a large part of the elected political class in recent years, and even so, there are few concrete measures to face the problem. In Brazil, the government supported no compelling anti-corruption reform agenda, and serious institutional setbacks took place. Furthermore, the advance of authoritarianism, with attacks on journalists and civil society, represents another threat to the fight against corruption. In 2020, with the rise of the Covid-19 pandemic, the risk of corruption increased. Countless corruption cases worldwide showed their most perverse consequences amid the terrible humanitarian crisis (IPC, 2020).
All these issues directly impact the Corruption Perception Index (CPI), an index in which, in 2020, Brazil was stagnant (considering the research's margin of error) at an appalling level, below the average of the BRICS, below the regional average for Latin America and the Caribbean and below the world average and still far from the average of the G20 and OECD countries (IPC, 2020). Data from the IPC (2019) show that corruption is more frequent in countries with power and money, especially in electoral campaigns.
When conceptualizing the phenomenon of corruption, it can be observed that it does not have a single concept, considering that it can be analyzed under different perspectives: social, economic, or historical, or under the prism of law, in the criminal sphere, civil and administrative (Gabardo & Reis, 2011). In this article, corruption refers to the "abusive use of a resource, power or position inherent to its functions by a public or private agent. e practice of any act or omission implying the violation of duties is also considered corruption" (Cunha, 2014, p. 45). Corruption is thus closely linked to political representation and delegation of power to elected politicians in democratic systems (Batista, 2013).
us, the spread of corruption has turned the issue into one of the main agendas of democratic countries (Batista, 2013). Corruption has also raised increasing interest from researchers in different areas of knowledge (Modesto, 2018). Most studies on the subject have sought to analyze possible articulations and motivations between deviant behavior and the occurrence of corrupt practices in public sector organizations (Carlos, Ribeiro, & Cintra, 2020;Macedo & Valadares, 2020;Medeiros & Forte, 2019;Medeiros & Freitas, 2019;Sodré, 2017;Souza et al., 2019). In some way, these studies show unethical and corrupt motivations of public agents in common and bring possibilities to reduce these practices, as is the example of transparency mechanisms.
A field of research has been established by Batista (2013) and Rodrigues and Rodrigues (2020), Andrade (2018), and Albuquerque and Ramos (2006) by looking at politicians and the effects of corruption on electoral succession. ese surveys found that factors such as reelection (mayors in their second term commit more corrupt acts than in the first term) and victory margin (the more significant the mayor's margin of victory (the less competitive the election) increase corruption (Batista, 2013). Andrade (2018), in turn, shows that the determining variables for the success of reelection are: transfers from the federal government, the mayor's political support base (share vote), and the existence or not of an episode involving the diversion of public resources. Mayors who adopt good governance practices tend to be re-elected (Rodrigues & Rodrigues, 2020).
In other spheres, Luciano, Wiedenhö, and Callegaro (2019) identified that a country that is less bureaucratic, more democratic, more transparent, and where institutions work better is more protected from the occurrence of corruption, considering that these four variables act as a deterrent to corruption. Araújo et al. (2020) present a theoretical discussion about the relationship between corruption in accounting, the "Brazilian way," and its implications for companies -without, however, bringing empirical results.
When analyzing the existing literature, as far as is known, studies that deal with corruption do not specifically address political agents at the micro-level, that is, municipal legislators. e scope of corruption is due to the propensity and involvement of people with corrupted actions (individual behavior), which are directly influenced by their perspectives and by the social interaction in the environment in which they are inserted (Den Nieuwenboer & Kaptein, 2008;Tavits, 2010).
Even if municipal political agents are supervised by the Office of the Controller General of the Union (CGU), e probability of being audited should discourage them from committing irregularities (Gomes, 2013;Andrade, 2018). us, this article analyzes the propensity for corrupt decisions of pre-candidates for mayor and councilor, using the 2020 municipal elections in 16 municipalities in the Midwest of Santa Catarina as an analytical framework. erefore, this study contributes by not focusing on the actions of public bodies but the behavior of municipal legislators. is article is organized into seven sections. is first section sought to contextualize the importance of the topic, define it, point out available studies on the subject, the research gap, and how this article seeks to respond to this gap. en, the theoretical foundations are presented, covering the theme of corruption, deviant behavior, corrupt decisions, and corruption's intrinsic and extrinsic motivations. Aerward, the methodological path, results, and discussions are presented. Finally, limitations and recommendations for future research are presented.
2 Corruption, deviant behavior, and corrupt decisions: analysis from Agency Theory and Public Choice Theory In its definition, corruption refers to an act or effect of corrupting and should be treated as a constant concern for democracy. Corruption is closely related to ethical principles, which are defined as the rules that define behavior (Bartels, 1967), being developed from the development of cognitive morality, evaluating the behaviors of a system of norms (Blasi, 1980;Kohlberg, 1969) e people involved creating a justifiable explanation for such an act, highlighting how their good intentions minimize the harmful consequences of their choices (Rabl & Kuhlmann, 2009). us, "corrupt behavior is a consequence of failures that are generated in the intrinsic decision-making process of each person" (Bandura, 1986:1;Tenbrunsel & Messick, 2004:1) When corruption is analyzed from the individual's perspective, one can look for explanations of its motivations in the Agency eory, which is based on the seminal studies of Jensen and Meckling (1976) and Eisenhardt (1989). e Agency eory seeks to explain the relationship between two or more individuals, where "one of these two individuals is an agent of the other, called the principal," where the agent "commits to perform certain tasks for the principal and the principal commits themselves to remunerate the agent" (Hendriksen & Van Breda, 1999, p. 139). e theory argues that it is almost impossible to ensure that the agent (in the case of this research -political agents) makes the best decision from the principal's point of view (the population), as this agent always seeks to maximize its utility.
It is necessary to understand politics as a process of exchange. e agents of these exchanges -politicians and bureaucrats -must be analyzed as holders of a utilitarian behavior like what they would have if they had acted in the economic sphere (Araújo & Sanchez, 2005).
e Public Choice eory should be cited by incorporating this behavioral premise (self-interest) to explain the political process. e theory argues that the choice of public agents cannot be expected to be identified with the pursuit of the common good (Buchanan, 1985;Araújo & Sanchez, 2005).
Power acts by changing the behavior of individuals in different ways, interfering with their decisionmaking execution (Bendahan et al., 2017;Sturm & Antonakis, 2015). According to Da Matta (1982, p. 1), it is crucial to abandon linear perspectives to understand themes such as violence, crime, and deviance. It is impossible to identify a single cause or origin for the facts. Becker (2008) explains that deviation is not a simple quality, present in some types of behavior and absent in others. However, it is the product of a process that involves reactions and interpretations of other people to behavior that is in focus. e deviation comes from the interaction between the person who commits it and the one who reacts to it (Becker, 2008:27).
Corruption is, therefore, an agent that alters the development of a country, state, and municipality, affecting its social development and the quality of life of its population that resides there (Luciano, Wiedenho, & Callegaro, 2019). erefore, understanding corruption as an organizational phenomenon is essential. It is also essential to enter the sociological contexts to improve understanding from the perspective of individual interpretations of public managers (Medeiros & Freitas, 2019).
Against this backdrop, corruption can be understood as a form of culture by the people who practice it. At the same time when the norms to fight corruption are absorbed by society. It will be challenging to combat the motivations that lead people to be involved with corruption. at is, individuals who grow up in societies where corruption predominates will probably accept it more easily, taking the practice of it (Medeiros & Freitas, 2019).
Countless problems transpose the field of study on corruption, such as, for example, high-level members, tenders and contracts, public works, and electoral campaigns (Souza, Silva & Gomes, 2019). e same authors (2019) explain that in most studies, the focus of the authors is not necessarily related to the issue or object of corruption, but to the institutional changes that are caused, and the mechanisms related to combating, preventing, or tolerating it, regardless of the reform and its purposes.
Regarding the circumstances that lead agents to deviant behavior and corrupt decisions, Dorminey et al. (2012 apud Carlos, 2019) described that at least three factors are necessary for their involvement: motivation, opportunity, and rationalization. ese factors reduce the feeling of guilt and disapproval for the commission of such an act illegal. As the focus of this study is on propensity, which comes from motivations, the following section provides a theoretical framework on the subject.

Intrinsic and extrinsic motivations to corruption
Most studies seek to create mechanisms to combat corruption focus on creating controls, monitoring, and punishments as the primary measures. Punishment increases generate an increase in cooperative behavior in people and are, therefore, the most important formal mechanisms for preventing corruption (Frost & Tischer, 2014). On the other hand, several studies use control and sanctions as part of a package of measures that need to be taken together to effectively reduce corrupt behavior and not just put it to sleep (Donato, 2018;Lange, 2008;Úbeda & Guzmán,2010).
New studies show that intrinsic and extrinsic motivations related to deviant behavior can be subdivided into secondary dimensions: perception of unethical decision-making, self-control, moral identity, and culpability (Carlos, 2019).
Initially, lack of self-control has been indicated as a predictor of criminal behavior. Criminals with low control over their attitudes tend to focus on the immediate benefit generated by the commission of the criminal act, giving less importance to the future consequences of their actions (Mead et al., 2009). Individuals with a high propensity for guilt tend to value moral and ethical traits (Köbis et al., 2016). Also, according to Veloso Gouveia et al. (2013), the constancy of self-control would explain why antisocial behaviors are so immutable throughout life. ose individuals who commit a significant number of deviant behaviors during childhood are more likely to continue with such behaviors at other times in life.
Moral identity, in turn, refers to the intra-individual property, which directs cognitions to the process of self-regulation, moderating the censorship and behavior used in its decision-making process (Moore, 2008). Resende and Porto (2017, p. 75) explain that "stages develop morality throughout life in which men stop responding only according to the rules until they are sure about what is right or wrong. Moral identity is not stable; it changes insofar as there is a stimulus or not for this type of behavior." Not all moral behavior is intentional because, in many circumstances, this way of behaving is automatic; its honesty comes from a previous experience in which there is unconscious processing of information (Reynolds & Ceranic, 2007).
us, according to Aquino and Reed (2002), an individual's moral identity may be associated with certain attitudes, beliefs, and behaviors. e central argument is that morality is central to the subject despite each individual's variations in moral aspects. It enables action through responsibility and conscience (Blasi, 1984). e perceptions of unethical decision-making are justified as a motivator of corruption in the assumptions of Ferrell and Gresham (1985). ey found that the behavior is influenced when there is an opportunity to do something wrong and unethical. e factors influence ethical or unethical decision-making.
e feeling of guilt arises when a person performs or even intends to do something not ethical (Aquino & Medeiros, 2009). For the authors, objective guilt occurs when a law is broken, and the transgressor is found guilty, even if he does not feel guilty.
Finally, people with a greater propensity to morally disengage will make unethical decisions more efficiently. In this case, the reverse is true, so they usually coincide when criminal action is underway (Abraham, Suleeman, & Takwin, 2018;Moore, 2008). rough moral disengagement, individuals feel free to self-penalize and feel guilty when their behavior violates proper standards. erefore, they are more likely to make unethical decisions (Detert, Treviño & Sweitzer, 2008).

Methodological Path
We have used exploratory and descriptive research with quantitative methods for data analysis. Furthermore, the assumptions of the survey design were followed. As a characteristic of this method, the interest is to carry out quantitative information on a sample (Freitas et al., 2000). According to Hair et al. (2005), survey research collects primary data from a representative sample of individuals. For operationalization, a structured questionnaire was used as a data collection instrument. e collection instrument was organized into 55 observable variables and divided into six blocks, as explained below: 1.
For the Socioeconomic profile of respondents: eight open and closed questions, namely: gender, age, income, profession, education, marital status, religion, and ethnicity.

2.
Respondents' perceptions of unethical decision-making: four statements, using the ethical compliance scale proposed by Pelletier and Bligh (2006), with questions built from the Likert scale (that is, ranging from 1-I totally disagree to 5-totally agree). is model was developed by Rensis Likert (1932) to measure attitudes in the context of behavioral sciences. 3.
Individual self-control: six observable variables that include multiple factors of their own that together make a person tend (or not) to consent to delinquency (Veloso Gouveia et al., 2013). us, this scale was adapted from Veloso Gouveia et al. (2013), and the questions were measured using a Likert scale. 4.
For the pre-candidate's moral identity: eight statements adapted from Resende and Porto (2017) were used using a Likert scale.

5.
Culpability: eight statements adapted from the Aquino and Medeiros scale (2009) with questions measured using the Likert scale. 6.
For Moral Disengagement: eleven questions adapted from the scale of Detert, Trevino, and Sweitzer (2008), were used measured by a Likert scale.
e study population of this research refers to pre-candidates for the municipal election in 2020. We opted for a non-probabilistic sampling, for convenience, with a single cross-section, which offers an overview of a given moment according to Hair et al. (2005). In order to make the data collection feasible, considering that it took place at the time of the Covid-19 Pandemic, it was decided to geographically locate the research locus for 16 municipalities in the midwest of Santa Catarina, all located within a radius of 150 km from Caçador/ SC, namely: Arroio Trinta, Lebon Régis, Calmon, Canoinhas, Campos Novos, Curitibanos, Correia Pinto, Fraiburgo, Iomerê, Major Vieira, Ponte Alta do Norte, Salto Veloso, Timbó Grande, and Videira. is region was chosen for accessibility. e self-administered online questionnaire was sent using e-mail, WhatsApp, and other networks. e questionnaire was initially sent to the political party directories in each municipality, with the data found available on online websites. In addition, it was forwarded to political representatives known to the researchers, who forwarded them to their relationship networks. As the number of pre-candidates was unknown to obtain a sample, we worked from the number of respondents, with 169. However, aer cleaning the database (where incomplete, inconsistent, or critical questionnaires were excluded in the comments and questionnaires of non-pre-candidates), a final sample of 115 valid questionnaires was obtained. Data collection took place between the beginning of August and mid-September 2020 (the date the candidates were announced -they were no longer pre-candidates). Table 1 shows the profile of survey respondents.
As it can be seen, the sample is composed mainly of men (73.9%), married or in a stable relationship (68.7%), Catholics (68.8%), white (80.9%), with education divided between complete high school (54.8%), postgraduate studies at the specialization level and MBA (26.1%), and the average age of 43 years. ese data approximate the average profile of candidates in the 2020 election in Brazil. Data from the Superior Electoral Court, published by the G1 (2020), show that the average profile of the candidate in the elections is a man, 46 years old, black, married, and with high school education. e issue that differs in this sample is ethnicity, as our sample is predominantly white, which may be characteristic of the region used as the locus.
Data analysis was performed using statistical techniques, using the SPSS Statistics 21 Soware. Initially, the data entry matrix preparation was carried out, where missing values, outliers, and the normality of the data distribution were analyzed in the preparation step.
Aer adjusting the data's normality and scale reliability, a descriptive, univariate, and bivariate analyses were carried out. e descriptive and univariate analysis presents the mean, standard deviation, minimum and maximum values of the frequency of behaviors. e data's bivariate analysis, differences between groups (performed using the One-Way Anova test), and Pearson's correlation test are presented to measure the statistical relationship between the dimensions studied.

Results
In this section, we present the univariate analysis of the dimensions and the descriptive and univariate analyses. e mean, standard deviation, and frequency distribution by scale intensity level are included in the appendix. In analyzing the dimension perception of unethical decision-making, 79% of respondents believe that other politicians let themselves be carried away by whims, passions, or personal interests in dealing with the public or other colleagues (average 4.05). 73% believe that other politicians have deliberately damaged the reputation of other servants or citizens who depend on them. Respondents believe that other politicians treat public service users differently, according to the user's social or economic prestige (average 3.93). us, it appears that respondents perceive unethical decisions by 'other' politicians (average 3.83).
In the analysis of the self-control dimension, it is verified that, in general, the sample claims to have good self-control. On average, 20% of respondent's pre-candidates are at risk of having fun, and they believe that stimulation and adventure are more important to them than safety. ey also show that they have reasonable control over their moods and attitudes when they are angry. However, they say they get irritated with some ease. However, in general, it appears that the sample disagreed with the statements that incite deviant behavior.
Regarding the pre-candidate's moral identity, it was identified that most of the respondents' precandidates believe that they do things in their free time (hobbies and activities) that identify them as a person able to hold an elective position in their municipality. However, 60.9% are actively involved in activities that communicate interest in having an elective office. For moral identity, 71.3% said they would feel good about having an elective office, 66% strongly desire to have one elective office in their municipality, and 55.6% of respondents said that having an elective office is very important to them -an average of 3.53. e analysis of these variables shows that a large part of the respondent pre-candidate desire and demonstrate the desire to hold an elective position in their municipality.
It is also observed that 47% of respondents do not feel guilty for having bad thoughts, and 43.5% also do not feel guilty for having some feelings they would not like, with a large percentage, about 48.7% said they did not if they feel guilt or remorse for feeling jealous. It appears that, in general, respondents' culpability is low, except for variables that involve not having time for family (average 3.33) and not being able to fulfill their obligations (average 3.86).
Regarding the pre-candidate's moral disengagement, it appears that the respondents show a very low propensity (total average of 1.89). However, it is noteworthy that issues that attack family or friends show a slightly greater propensity of some respondents to moral disengagement; 40.9% of respondents said it is okay to make great efforts to protect their friends who got into trouble. Furthermore, 43.5% show that it is okay to attack someone who threatens a family's honor. Table 3 analyzes differences between groups. In Table 3, there is only one statistical difference between the groups: e individual self-control and moral disengagement of those who have already occupied electoral positions and those who have not. In this case, it appears that those who have already held an electoral post have more self-control but are more willing to disengage morally -in comparison with pre-candidates who have not held elective posts previously. ere was also a statistical difference between guilt and gender, with women presenting more guilt than men. Furthermore, there were no statistically significant differences between groups.
Aer this stage, a bivariate analysis was carried out (see table 4) to verify whether there is a correlation (positive or negative) between the variables that explain the propensity to corruption and the profile variables of the pre-candidates. For the analyses, Pearson's linear correlation coefficient was used, given the normality of the data. In the Pearson correlation coefficient (r), as shown in Hair et al. (2005), is considered a: (1) solid relationship when the variation is between 0.91 and 1.00; (2) high when the variation is between 0.71 and 0.90; (3) moderate when the variation is 0.41 and 0.70; (4) small but defined when it varies between 0.21 and 0.40; and (5) light, almost imperceptible when the variation is between 0.01 and 0.20. Aer analyzing the correlation data, it was possible to verify a significant and positive association between moral identity and moral disengagement. In other words, moral disengagement is positively related to the variable of holding a previous elective office or not. ere is also an association between gender and guilt.
6 DISCUSSIONS is section of the article is intended to discuss the results, that is, the main advances in the theoretical and practical field. e first discussion that deserves attention is the low participation/representation of women, including as pre-candidates -although it is not different from the average representation of candidates for elections at different political levels. e findings corroborate with the findings of Araújo (2012), who states that women face two disadvantages, based on the initial formal exclusion from citizenship (considering that women have recently obtained the right to vote and participate in politics) and on current exclusion (with participation still unequal with men). In particular, Mazotte (2020) affirmed that the presence of women in the legislature is the lowest among the three branches. e author explains that the female presence is still disproportionate, given that the party defines who will have the most prominent among the candidates, television time, and of course, funding. Nevertheless, according to Araújo (2012), women have difficulties occupying power positions, but when they are elected, there is difficulty in having an active voice in politics.
e second result that deserves attention is the sample's perception of unethical decision-making by politicians. e results show that the sample considers the 'other' politicians predisposed to unethical decision-making, but not themselves. is issue can be explained and corroborated by the findings of Abraham, Suleeman, and Takwin (2018). ey argue that in these cases, there is the deactivation of selfregulation. ey do not think about monitoring themselves, about controlling themselves to minimize the roles of an individual in the act of corruption. at is, deactivation consists of shiing or diffusing responsibility.
Regarding the self-control of the pre-candidate sample, they generally show a high capacity for self-control, especially the part of the sample that did not hold a previous elective position. e results indicated that individual self-control was positively correlated with culpability and moral disengagement. According to Veloso Gouveia et al. (2013), it is composed of a series of individual principles that make a person prone (or not) to give in to the "temptation" of the crime. It is acquired through socialization, especially the family, in the early stages of life. Although it is difficult to explain why the sample of respondents who have already held electoral positions is less self-controlled, it can be assumed that there may be particular security of action. Because of familiarity, the person can be more willing to assume risks and give in to moral deviations.
In a nutshell, previous studies have already identified that the variable reelection (mayors in their second term are more corrupt than in the first term) and margin of victory (the more significant the margin, the lower the election competitiveness increases corruption (Batista, 2013). Similarly, there is evidence that mayors who adopt good governance practices tend to be re-elected (Rodrigues & Rodrigues, 2020). Our study corroborates what was previously presented in the literature.
Concerning guilt, a result that deserves to be highlighted is related to gender. at is, the results of this article showed that women have more guilt than men. Freud (1930Freud ( /1974 conceived the feeling of guilt as the tension between the ego and the superego, whose function would be to curb the individual's desire for aggression. According to Aquino and Medeiros (2009), the feeling of guilt arises because a person intends or even thinks to do something bad that goes against their ethical and moral principles. Wright (1971) defined guilt as an intolerable affective, directly followed by disrespect, which persists until some balance is restored by reparation or confession and forgiveness. Based on these definitions, it is possible to affirm that women are less prone to decisions that violate ethical and moral precepts or even corruption in electoral positions.
Although the average is relatively low in the dimension of moral disengagement, more than 40% of respondents are predisposed to disengagement to protect their friends and family. Detert, Trevino, and Sweitzer (2008) claim that individuals with high moral disengagement use cognitive mechanisms to minimize unethical impacts (cognitive dissonance). Consequently, suspending the processes of selfregulation of moral behavior and making fewer complex decisions can be assumed that these individuals can easily take this deviation to other spheres and decisions.
In practical terms, the sample has a low propensity for deviant behavior, that is, corruption. When raising this analysis, which response to the main objective of this study, it is impossible not to have new inherent questions. Why is Brazil one of the most corrupt countries in the world? 7 Limitations and recommendations for future research Some limitations must be listed even if the research has achieved its objective. e first limitation is that the number of respondents (although considered an adequate sample) is still tiny. Low participation is related to the exact number of pre-candidates not being known, and there is no pre-registration of this information. us, the collection was based on political parties. Even though this survey did not ask the party itself, it may have generated some discomfort/fear in respondents due to the unknown topic. us, the total number of candidates was used. e questionnaire could reach an even more significant number of respondents if we knew all the pre-candidates. In addition, because the collection took place during the pandemic period, and a large part of the municipalities have broad limitations, the survey was only applied online (via WhatsApp and e-mail). Of the 25 municipalities where the survey was sent, only 16 responded. at is, 36% of the municipalities did not return the contact, being possible to obtain more respondents if the contact was made in person.
A non-probabilistic sample, which depends solely on the researcher's criteria with procedures that are much more critical about the validity of their results (Gil, 2008), is also a limitation to be noted.
However, the question of the research theme raises another limitation related to the authenticity of respondents concerning measurable variables that analyze guilt, moral disengagement, and self-control. Respondents could be afraid of being honest (unconsciously or consciously) about such variables.
Furthermore, as it is a fertile field for further research, we may consider solving our limitations and that this one has continuity with the candidates effectively elected in the election. Nevertheless, other forms of analysis of observable variables are recommended, primarily focus groups and experimental studies (to avoid response biases) and obtain a deeper perspective on these behaviors. us, if municipal legislators had little propensity for deviant behavior at the micro-level, a research question that deserves to be investigated is why Brazil is a country with such high levels of corruption. Not only but the results of this study also raise the possibility of conducting observational and experimental studies to analyze the propensity of both municipal legislators and political agents to deviant behavior.
Longitudinal studies can bring an approach over time based on evidence of corruption, seeking to correlate moments of more and less corruption with macro-environment variables (political, economic, legal, and technological) and individual variables of legislators -at different levels of public management.