Autonomia e discricionariedade do Ministério Público no Brasil

Autonomie et pouvoir discrétionnaire du Ministère Public au Brésil

Fábio Kerche

Translated by Thiago Gomide Nasser
Translation from DADOS – Revista de Ciências Sociais, v. 50, n. 2, pp. 259-279, 2007

ABSTRACT

This article deals with the autonomy, discretionary power, and wide range of attributions allotted to the Public Prosecutor's Office in Brazil by the 1988 Constitution. One of the objectives is to analyze whether this combination is not alien to the democratic principle that state institutions, and even those that do not choose their members by direct elections, should be accountable to the public for their acts. The author draws on the neo-institutionalist literature on bureaucracy to analyze the Public Prosecutor's Office in Brazil. The conclusion is that there has been a quasi-abdication by politicians in relation to the institution in question, a rare occurrence in Brazil and in other democracies.

Key words: Public Prosecutor's Office; accountability; democracy

RÉSUMÉ

Dans cet article, on examine l'autonomie, le pouvoir discrétionnaire et le large éventail d'attributions du Ministère Public au Brésil figurant dans la Constitution de 1988, afin de savoir si cet assemblage ne s'oppose pas au principe démocratique selon lequel les institutions d'État, y compris celles qui n'utilisent pas le système d'élections directes pour l'admission de leurs membres, doivent rendre compte de leurs actes devant le peuple souverain. Dans ce but, on part de la littérature néo-institutionnelle concernant la bureaucratie afin d'étudier le Ministère Public au Brésil. En conclusion, on voit que les hommes politiques se servent très peu de l'institution, ce qui est plutôt rare au Brésil ainsi que chez d'autres pays démocratiques.

Mots-clé: Ministère Public; accountability; démocratie

Until the democratic 1988 Constitution, the Brazilian Public Prosecutor's Office - PPO- (Ministério Público) was an institution attached to the Executive branch of power – the prevalent model in most consolidated democracies. Although this agency has already occupied different sections and chapters in other constitutions during the country's republican history, institutional instruments, such as the appointing and dismissal of the attorney-general by the president, are proof of its attachment to the government. The 1987/8 constitutional assemblies which, however, decided to grant greater autonomy to the institution not only from a formal point of view, by including it in a chapter apart from the Executive branch, but also by creating mechanisms which considerably protect the state and federal-level PPO from governmental interference in particular and political interference in general. This autonomy, though, was not accompanied by a significant increase of instruments of accountability.

Autonomy is not the sole factor responsible for transforming public prosecutors into important political actors. Autonomy without instruments of action would not have been enough to catapult the PPO from the criminal section of the newspaper to the political section. The opposite is also true: granting the agency instruments of action – such as the public penal actions, public civil actions and civil inquiries – yet, without conferring it autonomy would transform the Brazilian PPO into an institution subordinated to the Executive branch limited to carrying out governmental decisions and guidelines.

Furthermore, the PPO has a wide range of attributions, warranting the claim that there are few issues in Brazilian society than cannot be transformed into a legal matter by this institution. From petty crimes to corruption, from river pollution to the right to stand for election, almost any topic can become judicialized at the will of this institution.

This article discusses the new PPO that arose after the 1988 Constitution. It concludes that the PPO is unique because it combines a set of traits – autonomy, instruments of action, discretionary powers, in addition to a wide range of attributions – that are not common in institutions submitted to few accountability mechanisms.

THE AUTONOMY OF THE PUBLIC PROSECUTOR'S OFFICE

Members of the constitutional assembly of 1987/8 detached the PPO from the Executive without subordinating it to the Legislative or Judiciary branches. Autonomy in relation to the branches of government, however, need not imply freedom from any kind of political accountability. The crux of the question consists of finding out whether there are instruments capable of ensuring that the action of public prosecutors is accounted for, whether they are responsive to an external actor, or whether members of the PPO can be held responsible for their actions when misconduct is identified.

The number of sanctions applied by politicians, the number of parliamentary commissions responsible for oversight, the requirement of annual reports of activities etc. are the clearest indicators in evaluating the degree of external interference over an organization.

If the only parameter were this kind of a posteriori oversight – which has been dubbed elsewhere "police patrol" oversight (Kiewiet and McCubbins, 1991) –

An alternative consists of seeking indirect instruments that could serve as incentives for the institutions to closely follow the will of politicians. A posteriori oversight of the police-patrol type is only one among many possibilities – all of them less efficient, it must be said – yet not the only one. That politicians do not constantly sanction members of the PPO and do not directly oversee their actions is not tantamount to claiming that public prosecutors do not take into consideration the will of politicians and that they cannot redirect the PPO's activity, if not immediately, in due time. In fact, this lack of sanctioning might indicate that prosecutors are anticipating what politicians desire so as to avoid possible sanctioning. Stated otherwise:

"[…] the fact that bureaucratic agents appear to make policy with little direct input from elected officials does not necessarily imply that bureaucrats are responsible for policy choices or that they employ meaningful ‘discretion.' Bureaucratic choice is embedded in a game in which the appointment power of the executive and legislature, together with the threat of sanctions, provides a potentially decisive influence over policy" (Calvert, McCubbins and Weingast, 1989:589, emphasis added).

The delegation of tasks and powers by politicians to a government institution, therefore, is not always synonymous with abdication. This implies that the "agent has complete discretion over the policy choices and that the principal has no control" (McCubbins and Noble, 1995:74). As the authors themselves admit, this definition is rather extreme, as there might be relative amounts of abdication "but relative amounts of abdication imply that the principal is able to influence the agent's choices to at least some extent" (ibidem). Thus, absolute abdication only exists between politicians and bureaucrats when the former have no means to modify the actions and initiatives of non-elected actors. When there is no kind of budget oversight and/or politicians possess the institutional instruments – even if they are only indirect means of control – it is not possible to define the delegation of tasks as a form of abdication.

According to this definition, bodies with a high degree of autonomy relative to politicians would not be truly autonomous. Ultimately, budget approval remains a prerogative of elected actors, even when such autonomous bodies propose their own budgets. According to this rationale, the PPO established by the 1988 Constitution bears great resemblance to the PPO during the military regime since, in both cases, budgets were ultimately approved by politicians. Furthermore, another possible argument is that, since legislative initiative depends on elected actors, politicians would be able to modify legislation in order to change the direction of a public agency, even in cases when the budget is managed with a relative amount of autonomy or in cases when an agency is bound to certain budget constraints. This definition, however, does not take into account that a constitutionally defined organization such as the PPO enjoys greater protection from outside interference than organizations defined merely by ordinary law. For instance, an agency that must regularly account for its actions to the Legislative branch is less autonomous than an agency which requires its principal to amend the constitution in order to effect a change in the "contract" which defines this agency's functions. However, a public agency which submits its budget to the Legislative branch is more autonomous compared to those who do not have this right. Without identifying these distinctions, one might conclude that no agencies detain autonomy and that the changes brought about by the 1988 Constitution with regard to the PPO are thus irrelevant.

In this article, I suggest that the concept of abdication is more useful when it is ascribed greater flexibility and considered relative to other organizations, both national and international, or when the approval of the budget by the Legislative branch is subjected to restrictions with the consequence of limiting political interference. The budget of the PPO, for example, is proposed by the PPO itself and submitted to the Legislative branch and the PPO is responsible for its own administration, which makes it stand out in relation to other government organizations and to the PPO before the 1988 Constitution. Furthermore, given that public penal action is an initiative exclusive to the PPO, constraining its budget can imply paralyzing the fundamental activity of law enforcement, since no other entity can legally fulfill this function.

In other words, the assessment of mere delegation is oversimplifying as it does not take into account the creation of a series of instruments, many of them constitutionally based, which hinder government or legislative interference in an organization. Yet, to state that abdication has occurred does not explain that the PPO budget independence, for example, is only relative. Thus, somewhere between delegation and abdication, there is the possibility of an intermediary phenomenon: an elevated degree of autonomy, nevertheless, unaccompanied by substantial instruments of accountability. This article's hypothesis is that quasi-abdication is the term that best fits the case of the PPO after the 1988 Constitution.

Institutional Instruments Guaranteeing Autonomy

In light of the literature, this begs the question concerning which instruments are capable of altering the type of action of prosecutors in the Brazilian Justice System and what problems might arise.

The underlying assumption is that politicians can punish organizations not functioning appropriately and reward those which best fulfill their roles by granting greater financial support to the latter at the cost of the former, for example. However, even if competition among organizations occurs, the PPO can only be partially punished since there are limitations as to the extent its budget can be cut. In addition, in the case of public civil actions, although other actors can employ this instrument, only the PPO can resort to civil inquiries and penal public action (which many times serve as an accessory to public civil actions). In other words, although the PPO does not have a monopoly over several issues, it possesses privileged means in comparison to other actors, rendering competition unequal.

Another instrument used to elicit bureaucratic responsiveness is the establishment that one agency's actions can always be blocked by the actions of another, thereby guaranteeing institutional oversight (Kiewiet and McCubbins, 1991; Przeworski, 1998). The problem of multiple agents with veto-capacity of one agency's actions is that the greater the number of actors with the right to veto, the harder it becomes to modify the status quo; the greater the number of control mechanisms over an agent, the harder it becomes for it to make changes it was designed to implement: "Checks, then, inhibit the ability of agents to take actions that the principal considers undesirable, but necessarily retard agents from taking desirable actions as well […]"  (Kiewiet and McCubbins, 1991: 34).

An important aspect, however, is that if the Judiciary branch can be included as a body with the power of blocking the action of other actors, thus prompting responsiveness on the part of non-elected actors. Ultimately, the Judiciary branch does not respond directly to politicians, and therefore the issue at stake here is institutional oversight amongst agents in which the principal is made up of politicians. If the Judiciary branch were one of these organizations, there would be no truly autonomous government actor. Even regulation agencies, which enjoy a high level of autonomy, can have their actions reviewed by the Judiciary branch. In other words, the Judiciary is not an actor participating in institutional oversight stimulated by a principal made up of politicians.

Nonetheless, if the Judiciary branch is counted as one of the actors with veto-power over the PPO and responsible for the institutions accountability, would it be possible to state that prosecutor action is limited? The answer to this question is yes, at least in most cases. Ultimately, in Brazil, prosecutors are those responsible for proposing legal action, but it is the Judiciary that is responsible for adjudication. Yet, there are many forms of action in which the PPO does not depend on judges, even if later these actions can be questioned in court by those affected:

"Problems related to consumers' rights, the environment, the community are, more often than not, solved without having to resort to judicial proceedings that would submit them to the Judiciary. In fact, prosecutors (…) give priority to the solution through settlements agreed upon by litigious parties, administrative procedures, requisition of measures to public and private bodies and other extra-judicial instruments. One estimate is that 90% of all issues are solved without the need to involve the Judiciary" (Sadek, 2000: 28).

Furthermore, in matters directly related to the political game, the Judiciary's answer can come too late. The time it takes to process a judicial case makes it hard for a politician accused of misconduct to, for example, clear his reputation before an election. As in most cases, when there is nothing preventing prosecutors from taking a public stand concerning accused politicians, the press is used to raise suspicion (and often to serve as the trial) that can be damaging to politicians. Even judicial prosecution for accusations that do not hold is remote and never a political possibility, distancing this type of punishment from being a clear mechanism of accountability.  

In sum, the Judiciary branch is not a typical inducement mechanism used to sway prosecutors towards the wishes of politicians precisely because judges do not respond to politicians. The Judiciary branch also is not a reasonable parameter in itself to determine if an institution enjoys high doses of autonomy – ultimately, there is always the possibility of appealing to courts, which would lead to the conclusion that no government actor is autonomous. In addition, there are several initiatives by the PPO that circumvent judges and procedures whose reaction might not be timely from the perspective of the political-electoral game.

Another important instrument that serves politicians in influencing an agency is the right to appoint the head of the organization, posting someone with whom there are compatible interests. However, politicians must also be able to remove appointees who do not observe their wishes. The fear of punishment – of losing one's post – is fundamental in generating incentives so that the agent follows the wishes of the principal (Shapiro, 1997; Finn, 1993; Calvert, McCubbins and Weingast, 1989). In their quest to keep their posts, agents anticipate the wishes of politicians. This instrument is so important that, Wood and Waterman, in their study of US agencies, found out that "in five of the seven programs we examined, agency outputs shifted immediately after a change in agency leadership" (1991: 822).

However, two institutional mechanisms undermine this explanation. The first point is that, when observing the rules concerning the occupant's removal, the head is considerably protected from political interference. In addition to the two-year tenure, the attorney-general can only be removed by the President, upon previous authorization by an absolute majority in the Senate. If the removal of a Secretary, a decision belonging exclusively to the President, is already considered costly from the political point of view (alliances may be shattered, posts occupied by parties are lost etc.), with the need of Senate participation, this initiative can be extremely difficult. This protection is an exception within the Brazilian political system, both with regard to most posts in the government structure and to other constitutions, with regards to the PPO. Even in a compared perspective, in general, the head of the institution which holds the monopoly over penal action is appointed and removed at the exclusive will of the head of the Executive branch.¹

Protection against removal at the exclusive will of the President is a prerogative not even the head of the Central Bank or the directors of other important public entities enjoy, for example. The relationship with the principal is not only fragile, but is also based on two distinct principals since there is no guarantee the President has a majority in the Senate. According to Kiewiet and McCubbins (1991), multiple principals may not be capable of expressing a unique policy, making it harder to evaluate the actions of agents and giving bureaucrats a margin for maneuver, thus making it possible to act counter to the interest of principals.²  Ultimately, since there are two principals, it is not clear which one of them the agent must take into account given that their interests do not necessarily coincide.

Furthermore, another factor limits the importance of the ability to appoint the attorney-general. The institutional instruments available for the attorney-general to control the action of other members of the PPO are also limited when they are in the position of principal. The structure of the PPO is not that of a traditional hierarchy, as in most other government organization. Prosecutors have a considerable amount of autonomy with regard to the attorney-general, as do prosecutors in relation to attorney-generals of states. The promotion of a prosecutor, which could hypothetically serve as an important mechanism of incentive as to align the interests of prosecutors and the attorney-general, occurs independently with no regard to the wishes of the attorney-general, since they follow either the criterion of seniority (senior members have priority in promotion) or the decision taken by collegiate bodies of the PPO. In other words, even if the prosecutor's professional performance is not aligned with the attorney-general's interests, there are institutional mechanisms that guarantee career ascension. This model may stimulate low predictability of the PPO's actions and thwart the creation of a unified and coherent institutional policy defined by the attorney-general.

The possibility of being maintained in a leadership post, although limited due to the existence of multiple principals, could serve as an incentive for the attorney-general to take into consideration the wishes of politicians. According to the Constitution, the federal attorney-general  can be kept in the same post as many times as the Senate and the President find it convenient. Therefore, in the attempt to keep the post, the attorney-general would attend to the wishes of the principals in order to assure maintenance. It is the same logic of the accountability vote transposed to a non-direct vote system.³ However, if the President does not have a majority in the Senate, whom shall the attorney-general try to "please"? In case it is the President, the attorney-general can be vetoed by the Senate. In case it is the senators, the attorney-general might not be appointed by the President.

A similar scenario is the appointment and dismissal of state-level attorney-generals. According to institutional rules, members of the state PPO participate in a direct vote in order to select a list of three names to be presented to the state governor. As the federal attorney-general , the state attorney-general  has a two-year tenure, as dismissal can only occur with an absolute majority vote in the state Legislative, regardless of any governor interference.

In this case, therefore, the principals are multiple: other colleagues in the state PPO, who vote to form the list presenting three candidates; the governor, who chooses from among those on the list; and the state legislators who can choose to dismiss the state attorney-general . Unlike the federal attorney-general , the state attorney-general  can only be reappointed to the post once, making the process less predictable. In a direct election, the political party plays an important role for candidates who are presenting themselves to the electorate for the first time, or when a member is a candidate for the last time. Without parties, politicians exiting public life would have no incentive to pay attention to the wishes of electors, since they would not be disputing future elections nor transferring their political legacy to a party. If we apply this reasoning to the case of attorney-generals of states, who only have a limited number of terms (two in total) and no party-affiliation, they could easily become "uncontrollable" during their second tenure.

In other words, although the appointment of the head of an organization constitutes an important instrument in guaranteeing the influence of politicians over a  government enterprise, this cannot happen in the case of the PPO in Brazil. This is because, first, there is no unique principal. Second, the post is not strictly a political one. Third, the PPO is not organized in the traditional hierarchical manner, that is, leadership has rather limited internal powers.

Another instrument capable of generating stimulus for non-elected actors to act according to the wishes of politicians is the fire alarm (Kiewiet and McCubbins, 1991). In the same manner as a posteriori police-patrol style oversight, it is not capable of guaranteeing that an agent will report truthfully about actions. Therefore, the fire-alarm style oversight seeks to collect information from those who are served by the bureaucracy, namely, the citizens:

]]> "[…] since it is the bureaucracy that is in charge of serving citizens, they are the ones who have the best information concerning performance. In addition, if politicians were concerned with the well-being of citizens, then the interest of citizens coincide with theirs, who are the principals, and not with those of bureaucrats, who are the agents" (Przeworski, 1998:58).

The fire alarm, therefore, gives politicians a chance to modify the actions of agents when organized groups who orbit the bureaucracy sound the alarm, warning of eventual bureaucratic failings. The problem with this alarm is that, when groups are not organized, they cannot make themselves heard by politicians, or, using Moe's terminology (1984), the decibel-meter does not function properly. Another difficulty is that, if politicians do not possess institutional mechanisms to modify the actions of their agents, there is no way to rapidly change the bureaucracy's actions, not even by sounding the fire alarm.

The fire alarm would hypothetically be an important instrument in the oversight of the PPO since, for example, prosecutors are not required to present politicians with an account of their actions – even if this does not imply that the agents would fully disclose information to their principals. The problem, however, is that the actions of the PPO are not always oriented towards organized groups, but, rather, often to isolated individuals and scattered groups. Thus, the alarm might not be loud enough to be heard by politicians. In the event it is heard, what are the political instruments available to sanction institutions for their misconduct? The issue at stake is that the institutional mechanisms for overseeing an agency are not only imperfect, but also do not leave a great margin for punishment, thus undermining accountability.

A development that could have limited the autonomy of the PPO was the creation of the National Council for the Public Prosecutor's Office (Conselho Nacional do Ministério Público) in 2005. This organization is composed of the federal attorney-general , four members of the federal PPO, three members of the state PPO, two judges appointed by the Supreme Federal Court (Supremo Tribunal Federal) and another one by the High Court of Justice (Superior Tribunal de Justiça) and two lawyers appointed by the Brazilian Bar Association (Ordem dos Advogados do Brasil), and two other citizens noted for their knowledge of law appointed by the House of Representatives (Câmara dos Deputados) and the Senate. It is attributed with the tasks of overseeing the financial and administrative management of the PPO, controlling the activity of prosecutors, and choosing a national audit body. Although this initiative might suggest demanding greater accountability from the PPO, as well as demonstrating that politicians did not abdicate their right to legislate in the sense of influencing the actions of the PPO, it is worthwhile remembering that the members of this organization are law enforcers, that most of them are members of the PPO (even the prosecutor in charge of the audit body), and that only two of them are directly appointed by congressmen elected by the direct vote of citizens. There is no sufficient data for definitive claims concerning the Council. However, if the idea consisted of more effective and daily external oversight, the composition of the council with the majority of members belonging to the PPO itself could be considered enough to transcend corporative elements.

Thus, it is possible to conclude that if complete abdication did not occur (since politicians can still amend the Constitution, modify the constitutional law, or interfere in the budget proposed by the PPO ), there was at least a delegation which assured a wide margin of autonomy and a range of tasks seldom seen among government  bodies composed by non-elected members. In other words, what occurred was a quasi-abdication.

WIDE RANGE OF ATTRIBUTION AND INSTRUMENTS OF ACTION

This quasi­-abdication, although rather alien to the principle according to which in a democracy the people have the sovereign power and exercise it through elected representatives, is identifiable in a few certain cases and is therefore not exclusive to the PPO in Brazil. Some agencies, such as the US Federal Reserve, the institution of ombudsmen in Nordic countries, and Brazilian regulation agencies carry considerable autonomy in relation to politicians or society as a whole. In some cases, high doses of autonomy can even be, if not desirable from the perspective of a system based on popular sovereignty, at least acceptable: either because it guarantees the freedom of actors to act contrary to the temporary interests of political parties, or because they provide certain political compromises with credibility by not allowing them to seem like the direct consequence of the actions of a certain restricted political group, or because they decrease the transaction costs of the Legislative branch.

Nonetheless, not every quasi-abdication is equal in terms of the amplitude of the tasks transferred to non-elected actors. It can be said that the examples of agencies with massive amounts of autonomy are generally those which perform more specific and focused roles in their interferences in the political game, in society, in the economy, or in public policy. It is ultimately easier to create legal rules and institutional mechanisms for bureaucrats whose functions are well defined and who do not possess discretionary powers. In contrast, a broader range of attributions and vaguer legislation - increasing the chance for a non-elected actor to abuse discretionary powers – implies the need for stronger accountability in order for the principal to follow the performance of the agent. According to Shapiro: "It is one thing, however, to place a policy beyond democratic control by the relatively fixed provisions of a constitution and quite another to place it in the hands of an agency of government wielding ongoing discretion" (1997: 289). Therefore, when the degree of discretionary power is limited, the chances of arbitrary behavior are also reduced, making the existence of agencies with high doses of autonomy more reasonable. On the other hand, the more freedom a government actor has the greater the oversight over its actions should be.

Building upon the observation that the Brazilian PPO has undergone a process of quasi-abdication, would it be possible to say that these normative recommendations relative to the limitation of tasks and discretion are being followed? The answer varies according to the function being observed. As is well known, the Public PPO is actually a single institution which, nonetheless, performs a multitude of tasks. Regarding the PPO "classic" task of bearing the responsibility to propose public penal action for common crimes, it can be said, as I shall demonstrate, that the degree of discretion is rather low. However, regarding its two other main tasks – the oversight of politicians and bureaucrats (a role similar to that of ombudsmen) and the oversight of the government's and citizens' legal compliance – such recommendations are generally not followed, ensuring prosecutors with plenty leeway for discretionary action without accountability.

When a crime occurs – a robbery or a murder, for example – the police force is responsible for investigation. Under the supervision of the delegado, the chief of a police precinct, a police inquiry is elaborated and submitted to a judge, who distributes it to a prosecutor. Based on this piece, the prosecutor will then bring the case to the Judiciary power, which will then produce a conviction or acquittal sentence.

What obligates the prosecutor to submit all cases to the judiciary, regardless of the seriousness of the case, is the so-called principle of legality, a model which has not been adopted by all countries. In the United States, for example, the district attorney has the choice of negotiating with the defendant. The defendant can, for example, be offered a sentence reduction in exchange for turning in accomplices. Prosecutors are allowed to do this, with no judiciary interference, based on the principle of opportunity. However, in 45 of 50 of its states, district attorneys are directly elected by popular vote, a clear mechanism of vertical accountability. In other countries in which discretionary powers are guaranteed to prosecutors, in general the agency bearing responsibility for penal action is connected to the Ministry of Justice. In these cases, it is usually the minister who makes the appointments for key posts in the agency, traces collective strategy and controls the sanction mechanisms for those who stray from guidelines, thus creating a unified and coherent institutional policy by means of the adoption of a very explicit mechanism of horizontal accountability. In the Brazilian case, regarding penal action for common crimes, prosecutors do not have discretionary power to decide whether or not legal action should be brought before the Judiciary branch, compensating, to some extent, the fragile instruments of accountability to which they are submitted.

Thus, the rule seems to also apply to the Brazilian PPO: less accountability implies less discretionary power; that is, in countries which follow the opportunity principle this relationship is inverted – more discretionary power is combined with greater accountability. In this specific case, quasi-abdication of control over the PPO in Brazil is less alien to democracy since it guarantees less discretionary powers to prosecutors concerning penal action for common crimes.

Prosecuting Politicians (or not): An Agency of Accountability

The existence of an agency responsible for overseeing politicians, such as the PPO, is admission that the classic instrument of checks and balances – branches of government limiting each other branches – alone is not sufficient. The complexity of the contemporary government has given rise to an array of specialized and auxiliary agencies which do not conform to the classic liberal model of democracy intended to keep in check the power of government actors – such as the auditor's office (ouvidoria) in Latin countries or the ombudsmen in Nordic countries, for example. Therefore, an agency responsible for overseeing politicians is not unique to Brazil. The difference lies in, among other features, the amount of independence of all of its members and the wide range of attributions. 

Yet, this type of agency is also a tacit acknowledgment that the elector's vote is a weak mechanism considering the level of complexity of the political game. As noted by Przeworski, Stokes and Manin (1999), oversight requires such an elevated amount of information that, without the aid of accountability agencies, the voter would not be able to control politicians. However, a distinction must be made between agencies that contribute to increasing the amount of information available to voters and those who also detain the instruments to prosecute politicians, as well. In other words, there is a significant difference between the ombudsman and the PPO in Brazil. Whereas the former collects information for voters or for agencies within the Executive branch which might act judicially, the PPO in Brazil has the right to bring politicians to trial with practically no prior need to consult another political actor.

The importance and necessity of agencies serving as instruments for submitting politicians to accountability is not capable of, by itself, avoiding criticism. The investigation of public figures can also be conducted so as to favor allies and tarnish rivals, especially considering the amount of discretionary power of the PPO concerning civil legal action. Although politicians participate in investigation procedures, such as the Parliamentary Investigation Committees (Comissões Parlamentares de Inquérito – CPI), prosecutors do not have the duty to prosecute those indicated by congressmen. The result is, therefore, a configuration that is rather strange in democracies: an agency submitted to a few mechanisms of accountability, yet in possession of a considerable amount of discretionary power.

Some counter-arguments to this critique point to numbers concerning actions against politicians presented by state PPO. Considering the 645 municipalities in the state of São Paulo, for example, by the year 2000, in 38% of them a mayor had been prosecuted (Arantes, 2002), a telling number. However, another dimension must be pointed out: if the PPO has considerable discretionary power to opt for civil legal action, in other words, if the agency chose the cases in which it prosecuted more than 200 mayors does this imply that mayors who were not prosecuted are innocent? What are the conclusions when a mayor is not prosecuted? It is possible to reach any of the following conclusions: by not prosecuting, the PPO is giving the mayor a clean slate certificate; the prosecutor has given priority to some cases and set less important ones aside, or; finally, the prosecutor in a certain municipal jurisdiction is of the more bureaucratic type and does not want to cause a stir. The criteria adopted by the members of the PPO who have discretionary power may, therefore, not be clear. As William West notes, in his citation of Kenneth Davis, "[o]ften the most important discretionary decisions are the negative ones, such as not to initiate, not to investigate, not to prosecute, not to deal, and the negative decisions usually mean a final disposition" (West, 1995:25).

Finally, another point must be stressed. As convictions for corruption in the Judiciary branch can be complex, it seems that taking an extra-judiciary route constitutes a kind of strategy for the Public Prosecutor's Office or, at least, for part of it. When the Office informs the press that it is investigating a certain politician, it might be contributing to a public opinion trial in which the liberal principle that all are innocent until proven guilty is not necessarily obeyed. Imagine this hypothetical extreme situation: one week before elections, a prosecutor announces to the press that a certain candidate is suspected of wrongly directing public funds when in public office. What are the odds this candidate will get elected?

Another important attribution of Brazilian prosecutors concerns overseeing compliance with the law, including constitutional law. The main instrument to this end is civil public action and the civil inquiry procedure.

Public civil action is a legal procedure which allows collective, diffuse and homogenous individual interests to be brought to Justice.⁴ The Union, states, municipalities, autarkies, government-owned enterprises, foundations, and associations at least one year old established to defend causes such as the environment, consumers' rights, and cultural and historical heritage patrimony, as well as the PPO, can resort to this mechanism. For this reason, even if the constitutional stipulation for civil action is located in the section concerning the PPO, it is not the monopoly of the institution, although the PPO presents 90% of them, according to Ada Pelegrine Grinover (Sadek, 1997). Therefore, the PPO is a privileged actor in the use of this instrument which allows a wide array of issues to become judicialized and guarantees the discretionary powers of the members of this institution.

Discretionary powers are further strengthened as a result of the PPO monopoly over the civil inquiry procedure – an instrument in the preliminary stage of judicial processing. The civil inquiry allows investigations to be conducted and coordinated by prosecutors, in that they can independently decide whether a case merits being transformed into public civil action. In other words, despite not having a monopoly over public civil action, the PPO is by and large the main actor in using it and still has a large amount of discretionary power in doing so. Therefore, one of the elements which formally served as a defense for the accusation of a lack of accountability (the obligation to bring the case to Justice, as in the case of penal action, would consequently curtail its discretionary power) loses its strength, reserving an agency with no elected members and low accountability with the role of deciding whether a case merits being presented to Justice.

The argument that the defense of certain interests by prosecutors by means of the public civil action coincides with the citizen's will does not mean that they exert any control over the institution. As stated by Gruber, "[i]f by happy coincidence bureaucrats act the way the citizens want them to, bureaucracy may seem to be less of a problem, but it is not under democratic control." This is because

"Control may occur through a process of anticipated reactions. If bureaucrats accurately anticipate what the hand of the citizen would do, and then feel constrained to act on the basis of that anticipation, a form of democratic control has occurred. If bureaucrats are wrong in their anticipation and act in ways the citizenry or legislature does not approve of, however, it cannot be said that their actions have been controlled by the citizenry" (1987:12-13).

It can be argued that prosecutors, concerning their role in initiating public civil action, merely abide by the law and, therefore, would not be interfering motivated by their political preferences, but rather, in order to act in accordance with constitutional precepts – especially in a country in which society is allegedly incapable of defending its own rights. In this sense, a constitutionally-defined optimal policy would be necessary, which would in turn justify the existence of an agency composed of non-elected members, with broad powers and independence from the political game.

Although constitutional principles can be defined as goals to be pursued, reaching them and defining priorities are debatable questions and are, therefore, subject to the discretion of politicians – and not of bureaucrats. In other words, if the logic of elections were taken to an extreme, every politician would try to maximize social gains through public policies, not motivated by altruism, but rather because pleasing all electors would increase the chance of getting reelected. However, there are budget limitations that thwart the complete realization of this rational option, a dimension that is not necessarily part of the concerns of the Public Prosecutor's Office. In other words, in democracies, choosing priorities is a task that belongs to elected politicians. When this choice is made not as a result of political confrontation but as result of a technical or legal decision, an important dimension of participation and popular interference is jettisoned. Thus, the judicialization of politics – the transformation of issues traditionally dealt with by the Executive and Legislative branches into legal action – goes hand in hand with the discourse that attempts to negate politics by casting a shadow of suspicion over political parties and politicians and which deposits its faith in the technicians in central banks, regulation agencies and all other sorts of institutions far from the reach of popular sovereignty.

Only by ignoring important institutional mechanisms introduced by the 1988 Constitution is it possible to claim that politicians abdicated control over the PPO. Some indirect mechanisms were kept, faithful to the logic of checks and balances in democratic countries. However, the new PPO is reasonably well protected from everyday political interferences. The conclusion is that we are facing a case of political quasi-abdication, which can be considered rare when compared to other Brazilian public agencies.

Within a comparative perspective, this quasi­-abdication, in itself, is not a unique phenomenon. It is not hard to find examples of institutions reasonably well protected from political interference. What distinguishes the Brazilian PPO is that the idealizers of the constitution, in addition to autonomy, granted its non-elected members considerable discretionary power. This is why the argument which states that prosecutors are strictly abiding by the law in their actions – and are thereby exempt from accountability mechanisms – cannot be sustained. Discretionary power, combined with autonomy and the possession of a wide range of attributions, makes the PPO in Brazil a rare occurrence in democracies.

REFERENCES

ARANTES, Rogério. (1999), "Direito e Política: O Ministério Público e a Defesa dos Direitos Coletivos". Revista Brasileira de Ciências Sociais, vol. 14, nº 39, pp. 83-102.

___. (2002), Ministério Público e Política no Brasil. São Paulo, Educ/Sumaré/Fapesp.

CALVERT, Randall L., McCUBBINS, Mathew D. and WEINGAST, Barry R. (1989), "A Theory of Political Control and Agency Discretion". American Journal of Political Science, vol. 33, nº 3, pp. 588-611.

GRUBER, Judith E. (1987), Controlling Bureaucracies: Dilemmas in Democratic Governance. Berkeley/Los Angeles/London, University of California Press.

KERCHE, Fábio. (2005), "Agências Responsáveis pela Ação Penal: Um Estudo Comparativo".  Revista Brasileira de Ciências Criminais, nº 55, pp. 127-152.

KIEWIET, D. Roderick and McCUBBINS, Mathew D. (1991), The Logic of Delegation: Congressional Parties and the Appropriations Process. Chicago, University of Chicago Press.

McCUBBINS, Mathew and NOBLE, Gregory. (1995), "Equilibrium Behavior and the Appearance of Power: Legislators, Bureaucrats and the Budget Process in the U.S. and Japan", in P. Cowhey e M. McCubbins (eds.), Structure and Policy in Japan and the United States. New York, Cambridge University Press.

MOE, Terry M. (1984), "The New Economics of Organization". American Journal of Political Science, vol. 28, nº 4, pp. 739-777.

PRZEWORSKI, Adam. (1998), "Sobre o Desenho Institucional do Estado: Uma Perspectiva Agent x Principal", in B. Pereira (org.), Reforma do Estado e Administração Pública Gerencial. Rio de Janeiro, Editora FGV.

SADEK, Maria Tereza (org.). (1997), O Ministério Público e a Justiça no Brasil. São Paulo, Idesp/Sumaré.

___.  (2000), "Cidadania e Ministério Público", in M. T. Sadek (org.), Justiça e Cidadania no Brasil. São Paulo, Sumaré.

SHAPIRO, Martin. (1997), "The Problems of Independent Agencies in the United States and the European Union". Journal of European Public Policy, vol. 4, nº 2, pp. 276-277.

WEST, William F. (1995), Controlling the Bureaucracy: Institutional Constraints in Theory and Practice. Armonk/London, M. E. Sharp.

WOOD, B. Dan and WATERMAN, Richard W. (1991), "The Dynamics of Political Control of the Bureaucracy". The American Political Science Review, vol. 85, nº 3, pp. 801-828.

Submitted for publication in April 2006
Final draft received July 2007

1 On a compared study of the Public Prosecutor's Office in Brazil and its counterparts in other democracies, see Kerche (2005).
2 This problem has been termed the Madison Dilemma (Kiewiet and McCubbins, 1991).
3 In the accountability vote, the elector punishes or rewards the candidate based on past performance. This is different from the mandate vote in which the elector selects the best proposal during a campaign. For more details, see Przeworski, Manin and Stokes (1999), especially the Introduction.
4 Diffuse and collective interests are "transindividual of an indivisible nature" (Arantes, 1999:88), and homogenous individual ones "derive from a common origin" (ibidem). The fact is that interests are broad ranging, and it is thus possible to judicialize almost any issue involving a large number of citizens. For more details, see Arantes (1999).