Insulated and undemocratic: the (im)possibility of social accountability in Brazilian courts of accounts

Th is article examines the transparency of the Brazilian Courts of Accounts to citizens, andwhether these Courts have acted in a way that facilitates social accountability over public administration. Th e multidimensional scaling technique was used to determine the integrity of the 32 Courts of Accounts in Brazil. Evidence suggests that Brazilian Courts of Accounts are still not transparent to society and count on few mechanisms to help and facilitate citizen control. Evidence also shows that the means of social participation is limited, and represented by the Ombudsman, the Transparency Web Portal, and the Citizen Assistance Service. However, this study identifi ed that the Fiscal Responsibility Law (LRF) and the Access to Information Law (LAI) had a strong infl uence on the Courts of Accounts. Although gradual, these elements can lead to future changes in terms of an increasingly prominent role of the Courts of Accounts as strong allies in the process of enhancing vertical accountability.


INTRODUCTION
Representation (democracy) is a mode of political activity which can activate various forms of control and supervision through citizen participation.It is not only limited to delegation (electoral democracy) but a political process that connects society and institutions.These continuous controls are part of a broader constitutive process of accountability which ensures that government officials (elected politicians or bureaucrats1 ) act in the interest of society (Urbinati, 2006).
Although the term accountability does not have a conventional translation in Portuguese, its meaning has been widely debated by authors such as O'Donnell (1998) and Przeworski, Stokes, and Manin (1999).In general, these authors establish that accountability is related to the obligation of representatives to report to the people, and is characterized by 3 stages: a) information (transparency); b) justification; and c) sanction.
Additionally, O'Donnell (1998) classified accountability in two forms, vertical and horizontal, where the former occurs through elections and social participation, and the latter occurs through institutional controls during mandates such as parliamentary control, judicial control, administrativeprocedural control and social accountability, among others.
Due to the Roman character of the Brazilian legal system, judicial and administrative-procedural controls are exercised by secular institutions, with predominantly bureaucratic characteristics, which limit the effectiveness of social accountability, inhibiting the democratic transformations projected by the original constituent in the Constitution of the Federative Republic of Brazil (CF, 1988) and distancing them from the effective sense of democratic representation.Despite these limits, such organs of horizontal accountability end up playing a relevant republican role since they function as checks and balances against other powers.
Courts of accounts (TCs) are institutions of administrative-procedural control which play an important republican role in the Brazilian national and sub-national context.However, when it comes to their democratic character, or more specifically democratic representation, there is a debate about institutional development and democratic transformations, which point out limitations in this process.Loureiro, Teixeira, and Prado (2008) and Rocha and Zuccolotto (2017) argue that the institutional development of TCs is incremental, being mainly influenced by external factors rather than transformations induced by the management practices.Additionally, Rocha (2017) observed that despite the institutional advances of TCs, they seem far from expectations of external control bodies when considering democratic representation.In addition, the authors highlight clientelism, patrimonialism, corporatism and bureaucratic insulation as some of the factors that contribute to the incremental development of these institutions, given the existence of resistance and resilience forces.This causes the persistence of conflict between bureaucratization and democratization.
In this way, TCs, given the power of resistance forces, can become a kind of panopticon (Bentham, 2008), that is, one who sees everything, but cannot be seen, directly contradicting the sense of democracy.Even though (Bentham, 2008) proposed periodic supervision by society as a solution, it does not subject the State and society to continuous interaction, as advocated by the concept of democratic representation.Thus, transparency and effective continuous supervision from society are essential conditions which prevent TCs from becoming increasingly self-absorbed and bureaucratized, distancing themselves from society and contradicting democratic precepts.
The control of public bureaucracy is essential for the proper functioning of democracies because according to Olivieri (2011), public bureaucracy is not just a set of public officials and administrative processes, but one of the foundations of the exercise of state power and democratic government.
Public bureaucracy has a lot of power and continues to seek more.To do so, it prefers to keep its knowledge and intentions secret.In this case, democratic power is the only way to counter bureaucracy (Weber, 2012).
In Brazil, seminal studies on bureaucracy focused on its relationship with the formation of the state and not with democracy (Olivieri, 2011).However, recent studies, such as that of Loureiro, Abrucio, and Pacheco (2010) and Cavalcante and Lotta (2015) have resumed the relationship between bureaucracy and democracy, seeking to understand its composition, functioning and relationship with elected politicians and democracy.
On one hand, TCs are fundamental to democracy.On the other hand, given their rational-legal authority, TCs can threaten democracy by concealing knowledge and intentions.Therefore, it is essential that TCs are transparent and have channels that allow society to exercise social accountability, strengthening democracy and subjecting bureaucracy to controls and accountability.
Based on this analytical theoretical framework, the general objective of this study was to analyze the permeability of Brazilian TCs to society, through 3 perspectives: a) transparency of portals; b) transparency of fiscal activity; and c) citizen participation.Permeability in this study refers to the practices of TCs which allow society to exercise controls, either through access to information provided or the existence of institutional spaces where organized civil society can participate in agenda proposal, inspection, controls and decisions.
In addition to this introduction, the article has 4 parts: a) a literature review to contextualize the discussion on the modest democratization of institutions of accountability in Brazil; b) methodology; c) analysis results; and d) conclusions and final considerations.

DEMOCRACY AND INSTITUITIONS OF ACCOUNTABILITY IN BRAZIL
In the 18th century, representative democracy was based on the sense of restraining democracy and building a limited and, consequently, responsible government.However, the idea of a representative government being inherently singular has produced two distinct schools of thought : electoral and representative models (Urbinati, 2006).
The early schools endorsed a view of representation that combines elitism in political institutions (domain of competence) and popular legitimation (domain of consent).In this context, representation is founded on the principle of division of labor and functional selection of expertise as expressed by actors such as Michels (1915), Mosca (1980), Pareto (1979), Schumpeter (1961), among others.
The second school of thought, as highlighted by Urbinati (2006), was explicitly democratic and aimed at avoiding the concentration of sources of legitimacy in state institutions and the reduction of popular consent to an act of authorization.In this view, representation is based on the theory of consent, where election is seen as an expression of the right to participate at some level in the creation of laws and not as a method of transferring people's idiosyncratic preferences to selected professionals and politicians.The main icons in this field of knowledge are Condorcet (1968) and Paine (2011).
Based on this insight, Manin (1997), Pitkin (1967) and Urbinati (2006) resolved the liberal concept of democracy but resumed the defense of representative democracy, arguing that it goes beyond electoral democracy, thus emphasizing that elections can produce a responsible government, but not necessarily a representative government.According to Urbinati (2006), representation (political) cannot be reduced to a contract (delegation) signed through elections, nor can it be reduced to the appointment of legislators as substitutes of absent sovereign, since its nature consists in being constantly recreated and dynamically linked to society.
The author reports that representation is a mode of governance which does not exclude participation, since the opposite of representation is not participation, but exclusion from representation.In addition, unlike the minimalist character of liberal democracy, representation aims to avoid the concentration of sources of legitimacy in state institutions and reduce popular consent in a single act of authorization, that is, during voting.
In this sense, Urbinati (2006) highlights representation as a form of political participation that can activate a variety of modes of control and supervision involving citizens.Thus, instead of a delegation scheme (electoral democracy), representation is a political process that connects society and institutions (the State), since in representative democracy the sovereign people delegate power to a representative, but also has the power to remove the same (negative power).
However, for this negative power to be exercised, institutional designs of democracies are fundamental, in other words, the institutional spaces where democracy is exercised defines a country's level of democracy (Bobbio, 2000).According to Bobbio (2000), the degree of democracy cannot be measured by the number of individuals who have the right to participate in public decision making, but in the number of spaces which allows the exercise of this right.
The existence of these spaces reinforces government controls, which, according to Arantes, Loureiro, Couto, and Teixeira (2010) constitute a crucial dimension in democratic order and must be exercised in relation to elected politicians and public bureaucracy.
The ways in which these controls are exercised are called accountability.In general, the term is understood as the obligation of a person, organization or group to account for their actions before a third party (authority, society etc.), justify them and suffer sanctions in case of misconduct, or receive compensation, if otherwise (Bovens, Goodin, & Schillemans, 2014).
Besides the concept, the forms of accountability are widely discussed in the literature by Przeworski et al. (1999) andSchedler (1999).In the present study, we adopted the classification proposed by O'Donnell (1998), which distinguishes accountability using two axes: a) vertical; and b) horizontal.The vertical axis proceeds through the electoral process and existing participatory bodies in the society, while the horizontal axis proceeds mainly through state institutions, whose traditional characteristics are still maintained in Brazil since their creation, especially bureaucratization and insulation.
The maintenance of these characteristics and the few institutional advancements observed among these bodies (Loureiro et al., 2008;Rocha & Zuccolotto, 2017) are partly due to the power of bureaucracy.Weber (2012) highlights its high potency and continuous efforts for power, being one of the main tools of bureaucracy, the concealment of knowledge and intentions.To the author, bureaucratic administration always excludes the public.
Accordingly, these institutional instances end up becoming less permeable to the mechanisms of vertical accountability, reproducing the panopticon model (Bentham, 2008), that is, one that sees everything, but is invisible.Bentham ( 2008), however, proposed subjection of the controller to regular inspections by society, a process which has not been observed in studies focused on horizontal accountability institutions in Brazil.
This type of inspection is limited because, in general, social accountability practices are regulated by bureaucratic bodies.According to Peruzzotti and Smulovitz (2002), social accountability is a non-electoral vertical control of political authorities which aims to monitor the behavior of public officials, expose and report illegal acts and activate the operations of horizontal control agencies.For the authors, social accountability is vital for democracy, as it goes beyond the moment of voting.
In Brazil, this type of accountability is still under development and has been expressed by organizations such as councils, participatory budgets, etc.More recently, a prominent mode of participation consists of practices of coproduction of control.Such studies are mainly based on the seminal work of Marschall (2004) and express new participation possibilities in the production, implementation and control of public policies and goods.
In this context, despite progress in the theoretical meaning of representation, the characteristics of the liberal democracy model become evident and seem rooted in the practice of Brazilian institutions, especially those of horizontal accountability, historically insulated and bureaucratized.
Among these horizontal control institutions, TCs stand out as institutionalized spaces whose objective is to ensure good governance of public resources.TCs emerged in Brazil during Monarchy to Republic transition, accompanied with the creation of the Federal Court of Accounts (TCU) in 1891 (Loureiro, Teixeira, & Moraes, 2009) and reinforcement and/or expansion of the CF (1988) and the Fiscal Responsibility Law (LRF).
Recently, under the support of democratic transformations, these TCs started to defend practices associated with the concept of democratic representation, such as transparency and social accountability.Although studies show advances in the competence of such control organs, very little progress has been made in institutional terms (Loureiro et al., 2009) despite incremental advances.
These authors further point out that resistance to institutional advances is significant, where the advances were mostly due to factors related to the critical juncture (mainly external pressures).Finally, they highlight striking features of clientelism and patrimonialism, in addition to bureaucratic insulation and low transparency.Rocha and Zuccolotto (2017) add that this resilience is strengthened by the predominance of category associations of the TCs, such as the Rui Barbosa Institute and the Association of Members of the Brazilian Courts of Auditors, which tend to act as elements of resistance to the progress of democratic organs.
Given the above framework, there is evidence that such institutions of horizontal accountability maintain characteristics of the Weberian bureaucracy, concentrating a lot of power.And contrary to the assumptions of visibility of representative democracy, TCs maintain their acts and intentions confidential and thus prevent control exercised by both elected politicians and sovereign people over acts practiced by the TCs.The possible existence of this bureaucratization compromises the circular relationship between the state and society and breaks the link between society and institutions, as highlighted by Urbinati (2006), compromising democracy.

METHODOLOGY
To achieve the proposed objectives, an exploratory research was conducted.Initially, an observation script was prepared for the Brazilian TCs, considering three analytical perspectives, in the words of Zuccolotto and Teixeira (2019): a) portal transparency; b) transparency of fiscal activity; and c) citizen participation.These perspectives were constructed considering the international standards and best practices formulated by the International Organization of Supreme Audit Institutions (INTOSAI, 1977) and The Organization of Latin American and Caribbean Supreme Audit Institutions (Organización Latinoamericana y del Caribe of Supreme Audit Institutions [OLACEFS], 2014), in addition to the International Budget PartnerShip (IBP, 2015) and the Brazilian rules that regulate fiscal transparency and access to public information -LRF and Access to Information Law (LAI).
Table 1 shows the perspectives and the number of items included.The observation form was composed of 78 items with yes or no options, where 35.9%, 53.8% and 10.3% of the items are related to citizen participation, transparency of fiscal activity and portal transparency, respectively.Portal transparency assesses, for example, the existence of a portal, its update period, navigability, interactivity and content.Transparency of fiscal activity assesses the compliance of items related to institutional information, the audit cycle, financial and budget information, in addition to information about ministers/ advisors.Citizen participation assesses aspects related to participation spaces in the process of choosing advisors, in the preparation of audit plans and in the monitoring of resultant recommendations/ determinations from audits.
Prior to the observations, the complete form was validated by specialists from Getulio Vargas Foundation (FGV) in São Paulo, TCU, Valladolid University, Spain, and the University of Minho, Portugal, for adherence of the questions to the research object.

Data collection and organization
In relation to the observation form, 78 questions were answered through direct consultation of the website of each of the 32 2 Brazilian CTs.During the consultation, all the observations deemed important were noted, such as difficulties encountered in accessing information, page either down or nonexistent, delay in accessing the webpage, etc.To guarantee the reliability of the data collection process, each page consulted was captured and stored with the aid of the Zotero software (Corporation for Digital Scholarship, n.d.), a personal research assistant.
Data collection regarding the first empirical phase was carried out between December 19, 2016 and January 30, 2017.All questions were structured in a similar way, that is, direct questions whose answer is dichotomous, limited to yes or no.
2 In October 2017, the Federal Supreme Court (STF) confirmed the decision of the Legislative Assembly of Ceará (AL-CE) to extinguish the Court of Accounts of the Municipalities of Ceará (TCM-CE).Thus, this state counts on the Court of Accounts of the State of Ceará (TCE-CE) to supervise the accounts of the state government and of all the municipalities of that federative unit.Thus, the number of courts of accounts in Brazil dropped to 33 (Globo News Portal [G1], 2017).The Court of Auditors of Roraima (TCE-RR) was also excluded, as it was not possible to obtain all necessary data.

Data analysis
In addition to descriptive data analysis, multidimensional scaling (MDS) was adopted, in order to understand the similarities and dissimilarities between the TCs.According to Marôco (2014), MDS is a multivariate exploratory technique that makes it possible to represent sparingly and in a reduced dimensional system (perceptual map), the proximity (similarities/ dissimilarities) between subjects or objects, from a set of multivariate attributes (measured or perceived).
The author further points that the MDS aims to represent a set of n objects based on n (n -1)/2 similarities or dissimilarities (proximities), in a reduced set of q ≤ n -1 dimensions, so that the proximities between objects in the new dimensional plane closely resemble the original proximities.
To elaborate the distance matrix between the objects (TCs), the Alternating Least Squares SCALing (ALSCAL) algorithm, or alternating least squares was utilized.The ALSCAL algorithm is applicable to data whose scale is at any level of measurement, and can be adopted in the case of metric or nonmetric analysis.Additionally, ALSCAL was the first generalized application algorithm and one of its main characteristics is the greater importance to major dissimilarities.
This characteristic is important for this study since for binary data and many TCs that do not present great dissimilarities, visualization on the spatial map is facilitated when priority is given to great dissimilarities (Marôco, 2014).
Using the ALSCAL algorithm, the quality of the non-metric MDS adjustments was evaluated by Stress measures, which reveal the degree at which the configuration proposed by the model represents the data set, and RSQ (square of the correlation coefficient between the distances in spatial configuration and dissimilarity -this measure has an interpretation similar to regression R 2 ), which demonstrates the explanatory power of the model (Kruskal, 1964).To calculate the distances between the TCs, the Euclidean distance model was adopted.

RESULTS AND ANALYSIS
4.1 Overview of the Brazilian courts of accounts after application of observation form.
As discussed in the methodology, the measure for assessing permeability of Brazilian TCs was developed based on 3 perspectives: a) citizen participation (Participate); b) transparency of fiscal activity (Fiscalize); and c) transparency of portals (Portal).Thus, Table 3 shows the percentage of items satisfied by each TC according to the methodology adopted in this research.Source: Elaborated by the authors.Note: TCs are organized by the regions of Brazil."TCE" means "State Court of Account" and is followed by the abbreviation of the state."TCM" means "Muncipal Court of Accounts"; when the letter "s" is added, it means "Court of Accounts of Municipalities", followed by the acronym of the entity of the Federation of origin."TCU" means "Federal Court of Accounts".And "TCDF" means "Federal District Court of Accounts".TCE-RR was excluded due to lack of data in the perspective studies during the analysis period.
From Table 3, only 4 of the 32 TCs showed a compliance greater than 50% of the total items analyzed (78 items): a) TCMs-PA; b) TCE-TO; c) TCE-PR; and d) TCU.Two TCs showed a 50% compliance (TCE-AC and TCE-SC) of the total of items analyzed, mainly related to financial and budget information and that of ministers/advisors, indicating mere compliance with the legislation.The other TCs showed a compliance below 50% of the total of items analyzed, indicating that the controllers do not seem to satisfy their own Brazilian standards, undertaking the role of panopticon, that is, it sees everything, but is not seen, therefore, controlled.
Regarding the perspective transparency of fiscal activity, two TCs stood out: a) TCU; and b) TCMs-PA.They showed compliance greater than 70% of the total items analyzed (42 items), followed by TCE-SC and TCE-AC, which exceeded 60% compliance in relation to the total items analyzed.Table 3 also shows that 13 TC equivalent to 40% of the analyzed TCs, presented compliance greater than 50% of the total of items analyzed (42 items).
It is also evident that the percentages of service are quite heterogeneous and may result from inherent audit practices, which are not standardized among Brazilian TCs, generating significant interpretative differences between the agencies.
Regarding portal transparency, TCs are still far from their expectations.Despite the distinction of TCE-RN and TCU, which showed 87.5% and 75% compliance, respectively, for the items analyzed (8 items), most TCs (78%) did not show compliance greater than 50%, confirming a low level of portal transparency.Given that it is up to these organs to ensure the transparency of sub-national entities, the observed example is alarming because in addition to not serving as a parameter of the powers they oversee, they prevent the legitimate holder of power (the people) from obtaining timely and reliable information to request justification and, if necessary, punish responsible agents, even if it is symbolical.Low transparency limits social participation and, in this perspective, TCE-MG and TCE-RO showed greater compliance with the items analyzed, with a service of 35.7% (of 28 items analyzed), followed by TCE-AC, TCE-ES and TCU, with 32.1%.Table 3 also shows that, of the 32 TCs analyzed, only the aforementioned ones presented a compliance greater than 30%, indicating low participation of civil society among these organs and the remote exercise of social accountability which, in turn, results in low democratization.
Although descriptive statistics allowed inferences to be made about the level of transparency and citizen participation in Brazilian TCs, these data do not allow the visualization of the proximity between them.For example, when we look at the perspective of citizen participation, TCE-AC, TCE-ES and TCU show 32.1% compliance.However, this does not mean that these TCs are similar, since there are multiple possible combinations of responses among the 28 items analyzed.
In order to facilitate the perception of similarities and dissimilarities between objects (TCs), the results of the MDS are presented below, according to each of the perspectives adopted in this work.

Perceptual map of portal transparency
In this perspective, which considers the evaluation of the TC portals, based on the existence of the portal, update period, navigability, interactivity and its content, we observed that many TCs are similar, given the great overlap of points in the perceptual maps.Despite the similarities, 5 well-defined groups of similar TCs are identified:  Source: Elaborated by the authors.Note: Some courts of accounts can overlap and may not be visible in the map.
Although TCE-RN and TCU orbit around the groups, they do not share the same group when considering two Euclidean distances.Th ese TCs showed a higher percentage of compliance in this analytical perspective, with 87.5% and 75% compliance, respectively.TCE-BA, TCE-GO and TCE-RO showed diff erent percentages of compliance for the items analyzed, but when considering similar items answered, they share the same group.
Despite the advances in relation to the transparency of portals of the Executive Branch, within the scope of the TCs, the results show that many of them do not meet the minimum requirements of international organizations as well as comply with the Brazilian legislation.In addition, the practices adopted for portal transparency are diff erent, making citizen access, social accountability, accountability of public agents and the democratization of these bodies diffi cult.
We found that when considering classic MDS with three dimensions, the model's verdict measures showed Stress values of 0.12794 and RSQ of 0.9185, indicating moderate and very good fi t qualities, respectively.

Perceptual map of transparency of fi scal activity
Th is perspective assesses the compliance of items related to the audit cycle, information about TCs and their ministers/advisors and fi nancial and budget information.
Following the pattern of the perspective previously analyzed, it showed great dispersion among the TCs, making it diffi cult to identify groups.When this is possible (eg, when TCMs-GO and TCE-PB are grouped), the next stage of grouping (Euclidean distance) is large, showing that the disclosure practices of audit activities are very dissimilar between the entities.As there are no standardized procedures for carrying out audits in the diff erent TCs of the Brazilian Federation, it is not surprising that the disclosure practices have no similarities.
It is also added that the TCE-SC and TCE-BA were the most dissimilar among all TCs and showed close similarities, indicating that, in addition to being the ones that satisfi ed more analytical items, the conforming analytical items were similar among them.Although not a cause, it should be noted that the existence of spaces for co-production of control in these TCs can favor proximity, since it induces the adoption of better practices.It is also noted that when considering the classic MDS, with 3 dimensions, the model's verdict measures showed Stress values of 0.13946 and RSQ of 0.8662, indicating moderate and good fi t qualities, respectively.

Perceptual map of citizen participation
Th e last perspective analyzed refers to citizen participation, which assesses the openness of TCs to the participation of society, whether in the selection of advisors, in the preparation and monitoring of audits, among others, and is illustrated in Figure 3. Before any analysis, it is highlighted that the dimensional model presented in Figure 3 met the verdict values for the MDS solution, presenting Stress values of 0.08733 and RSQ of 0.9592, considered by Kruskal (1964) to be good and very good, respectively, indicating that the reduced dimensional model presented is adequate and can be analyzed.
As shown in Figure 3, TCE-MG and TCE-RO are very similar, since the point that represents them on the dimensional scale is superimposed.Th is means that these 2 TCs, which satisfi ed 35.7% of the items analyzed in this analytical perspective, met items of similar analysis.TCE-AC and TCE-ES, despite having equal percentages of conformity (32.1%), do not belong to a common plane in the perceptual map, indicating that, despite being close, they did not satisfy the same analyzed items, different from that observed for TCE-ES and TCU.
From the Euclidean distance, the perceptual map identified 4 groups more similar to each other and, consequently, more dissimilar from the others, namely: • It should also be noted that TCE-MT, TC-GO and TCM-RJ showed the greatest dissimilarity and, therefore, ended up orbiting close to the groups, without taking part in any of them.
When the perceptual map is coupled with descriptive data analysis, it appears that, although many TCs have equal percentages of conformity as regards items analyzed, they are not necessarily equal, since the conforming analytical items may vary.Thus, we found that in addition to the lack of development of mechanisms for social participation, the few practices in this regard are different among the TCs, showing autonomy, but also the lack of coordination of procedures between among these TCs.
In short, with regard to citizen participation, it can be said that, in general, TCs have a low level of permeability to society, that is, they are still insulated and, consequently, less democratized.Although there is an increasing appeal for more spaces of participation, no Brazilian TC offers the opportunity or mechanism for effective participation of society: a) appointing ministers/advisors; b) preparation of audit plan or participation in audits; and c) overseeing the compliance of determinations and recommendations by public managers.In relation to channels or mechanisms of dialogue with society, they are basically restricted to the ombudsman.
It is observed that although the predominant discourse is to encourage social participation, the three perceptual maps show the incremental development of Brazilian TCs, corroborating the understanding of Loureiro et al. (2008) and Rocha and Zuccolotto (2017).Even though there is no legal impediment to promoting transparency and spaces for participation, there seems to be a strong resistance on the part of these organs, resulting in a modern panopticon, that is, the one who is authorized to see everything, but who does not like to be seen.The non-transparency of these institutions, according to Birkinshaw (2006), causes a triple impact in the construction of accountability: a) prevents or makes it difficult for citizens, individual or in groups, to exercise control over State institutions (over the power); b) does not allow civil society-State interaction for the co-production of goods, control and information; and c) mitigates the contribution that TCs can make to democratic improvement (Bovens et al., 2014;O'Donnell, 1998;Przeworski et al., 1999).

FINAL CONSIDERATIONS
Since the 19th century, Weber (2012) realized that the relationship between politics and bureaucracy was marked by constant tension because the technical knowledge of the latter allowed the accumulation of power, which could threaten democracy.Bureaucrats, within a pattern of rationallegal domination, would seek to increase their power, using secrecy of knowledge and intentions as a weapon.Bureaucracy breaks the circular relationship that connects the State and society in representative democracy, by concentrating power and not exposing itself to the controls of elected politicians and, above all, society.
In this perspective, Brazilian TCs due to their bureaucratic characteristics and secular composition, became the object of this study, whose objective was to analyze the permeability of Brazilian TCs to society, through the analysis of portal transparency, transparency of fiscal activity and existing mechanisms for citizen participation.Although the terms transparency and social accountability are evident in the speeches of TC ministers and advisors, our results indicate that they are still insulated, given that, in general, they are not very transparent and do not adopt practices to encourage social accountability.
In relation to citizen participation, it was found that TCs have a low level of permeability to society, that is, they are still insulated institutions and, consequently, less democratized.Although there is an increasing appeal for more spaces for participation, no Brazilian TC offers a mechanism for effective participation of citizens in: a) appointing ministers /advisors; b) preparing audit plans or participating in audits; and c) overseeing the compliance of determinations and recommendations by public managers.The channels or mechanisms of dialogue with society are basically restricted to the ombudsman.
Therefore, when Brazilian TCs need to expose themselves to society, they comply, although incompletely, with the normative prescriptions foreseen in the LRF and LAI.Therefore, in terms of accountability, the TCs still do not allow the broad and easy monitoring of their actions.This undoubtedly prevents, or at least makes monitoring of their behavior through social accountability a difficult task, and thus, delay, if not render unfeasible, actions that expose and denounce illegal acts practiced by them, or even activate operations of other horizontal control agencies, as provided by Peruzzotti and Smulovitz (2002).
Although the scenario is not favorable, incremental advances can be seen when it comes to mechanisms to encourage citizen participation.Based on the 32 TCs analyzed, 15 have a page, tab or guide on the internet dedicated specifically to citizenship, even though such spaces comprise of scattered and poorly structured information, requiring citizens or interested individuals to invest a reasonable amount of time to find such information.In addition, it is possible to observe the existence of initiatives, in some courts, related to the co-production of control, such as itinerant ombudsmen and public hearings.
Through the compliance script, it was also possible to identify that all TCs have an institutional teaching or training structure.This mechanism proves to be important, insofar as the TC could act by enabling citizens to carry out more effective control.However, not all TCs offer courses or lectures to society.In most cases, such courses are offered to train their own civil servants and public managers of the Executive Branch.However, 9 of the evaluated TCs offer training courses to public policy councils.
When it comes to channels or mechanisms for dialogue with society, the main means used by TCs are the ombudsman, physical and via the Internet, the transparency portal and the citizen information service.However, in 3 TCs we identified: a) public debate/dialogue; and b) the existence of an online radio or television channel.Despite this, in a few cases, the programs are used to convey and discuss the outcomes of actions undertaken by the TCs.In large, such programs are used almost exclusively to transmit the sessions performed by the TCs.
Regarding the disclosure of outcomes of actions undertaken by the TCs, it was found that, in general, the TCs do not disclose the accounts presented by public managers on their respective websites.Even among the few that do, accessing these accounts is not an easy task.This also applies to: a) the availability and access to prior opinion on government accounts; b) availability and access to the list of agents convicted of the misuse of public resources; and c) audit results.We also verified: a) incomplete information; b) low comprehension of available information; and c) restrictions as regards dissemination channels.
From the joint analysis of the 3 perspectives, it appears that Brazilian TCs, in general, are still far from the desired permeability required in order for social accountability to be more efficient in its supervisory role, limiting the continuous exercise of accountability beyond voting (Urbinati, 2006).It is also evident that the mechanisms that aid or facilitate the exercise of social control over public administration are weakened.This picture shows that, despite the institutional advances highlighted by Loureiro et al. (2008) and Rocha and Zuccolotto (2017) on democratization, there seems to be little political will, highlighting a threat to democracy in agreement with Weber (2012).In addition to concentrating a lot of power, such institutions prefer to keep their actions secret to avoid political or societal control.
Thus, TCs remain distant or even insulated from society which is not beneficial to both the institutions or society, since the possibility of reducing information asymmetry between the representatives and the represented ends up in a long horizon to be traced.This situation reinforces the need to review the institutional design of Brazilian TCs, so that besides being a horizontal accountability entity, it can be accountable, allowing parliamentary control over actions and, above all, the control of society.
However, it should be noted that, despite the unfavorable scenario when it comes to the democratization of TCs, bear in mind that in the 30 years of redemocratization, advances have occurred, mainly due to CF (1988), LRF and LAI, which are the cornerstones for the democratization of TCs, which also made this research possible.Of course, these institutions still have a long way to go, but the mere incorporation of transparency and social control into the discourse of such control institutions already creates a favorable environment for the implementation of actions that can result in democratic advances.
In addition, although there are forces of resistance in this public bureaucracy, many of the changes and advances that have occurred in the TCs stem from the very resilience on part of this bureaucracy (Loureiro et al., 2008), indicating that patrimonial and clientelist characteristics coexist with innovative and transformational forces of these structures.
Finally, it should be kept in mind that this research was limited to analyzing the portals of TCs and consequently, understanding existing spaces of social participation and accountability.For this reason, it is recommended that future research interview advisors and civil servants to understand how they interpret the need for democratic transformation of TCs and what actions have been taken, if any, in this respective.

TABLE 1 PERSPECTIVES
AND NUMBER OF ITEMS ANALYZED.
Source: Elaborated by the authors.