Latin American criminal justice no longer rejects alternatives to incarceration as proper responses to criminal activity. A rise in violent crime during the last democratic wave led governments across the region to adopt imprisonment as the default penalty for lawbreakers and to facilitate the use of pre-trial confinement. In contrast, legislation enacted since the mid-1990s has gradually expanded the possibility of alternative sentences for small crimes and imposed limits to the pre-trial detention of non-violent suspects.Footnote 1 These changes are clear reactions to some of the worst consequences of prison overcrowding, such as frequent and violent riots. Yet they are also puzzling because voter support for harsh penal laws continues to discourage politicians from weakening existing sentencing frameworks.Footnote 2 In the absence of electoral incentives for penal reform, what explains attempts to limit the state's punitive capacity?
This question is worth answering because scaling back incarceration rates can improve the quality of democracy. The average level of occupancy inside Latin America's penitentiaries has steadily increased since the 1980s and now stands at more than 140 per cent.Footnote 3 Overcrowding is associated with deteriorating living standards, violence and a host of human rights violations that highlight the limited reach of civil rights in the region.Footnote 4 Furthermore, high imprisonment rates can have a negative impact on the political and civic engagement of communities whose members represent a significant percentage of the prison inmate population.Footnote 5 A final, and related, point is that harsh sentences mostly affect the lower classes and racial minorities, undermining the building of a social order rooted in legal equality.Footnote 6 Sentencing reform will not solve all these problems, but to understand how it comes about is to uncover one way in which Latin American countries can inch closer to the liberal-democratic ideal that eludes them so far.
This article claims that the adoption of alternatives to incarceration is contingent on the presence of autonomous policy elites within the state. By this I mean bureaucratic actors who are insulated from networks of political patronage and who enjoy broad mandates to develop solutions to pressing social or economic problems. In the case of sentencing reform, bureaucrats seek to solve penal crises that threaten social stability, impose heavy fiscal costs and undermine public perceptions of state authority.Footnote 7 These policy elites develop new sentencing rules, use their expertise to influence congressional debates and shape the timing and scope of controversial projects to facilitate their approval despite popular opposition.Footnote 8 The argument suggests that the likelihood of reform increases when the requirements of governance do not force presidents to exchange bureaucratic posts for political support, and when public concern over crime does not become so heightened as to lead presidents to limit bureaucratic discretion.
The focus on autonomous bureaucratic actors fills an important gap in the knowledge of how criminal justice reforms come about.Footnote 9 Specifically, it refines and complements theories that emphasise the causal impact of political ideology and international activism on processes of policy change. I argue that the election of parties ideologically committed to civil rights protection and the formation of transnational networks of human rights activists facilitate reform only insofar as they provide autonomous bureaucrats with additional protection against political pressures. Without bureaucratic autonomy, however, parties that support sentencing reform are unlikely to overcome electoral barriers to change, and external activists are unlikely to gain access to the decision-making process.
To support this argument the article analyses the process behind changes to Brazil's sentencing framework. This is a fitting case study because the country represents a most-likely instance of bureaucracy-driven policy change: it has legislated for alternatives and limits to incarceration despite strong societal opposition to these measures. Brazil (a) broadened the applicability of non-prison sentences in 1998, (b) created alternatives to pre-trial confinement in 2011, and (c) is engaged in a debate about the reduction of penalties for street-level drug dealers. If autonomous policy elites are crucial agents of policy change in that country we should expect similar actors to play an important role in criminal justice reforms across the region as well.
The next section further explores the relationship between autonomous bureaucrats and sentencing reform. The article then analyses the political context of policy change in Brazil, identifies the country's main agents of reform, explores their level of autonomy and traces the process by which the projects mentioned above navigated the country's political system. The analysis covers three presidencies: that of Fernando Henrique Cardoso (1995–2002) of the Partido da Social Democracia Brasileira (Brazilian Social Democratic Party, PSDB), and those of Luiz Inácio ‘Lula’ da Silva (2003–10) and Dilma Rousseff (2011–present) of the Partido dos Trabalhadores (Workers’ Party, PT). The conclusion summarises the findings and calls for more research on the relationship between bureaucrats and criminal justice reform in Latin America.
Bureaucratic Autonomy and Sentencing Reform
Autonomous bureaucrats are politically insulated from patronage networks and granted broad mandates to pursue policy projects. Insulation means that office holders do not use bureaucratic positions as currency in the building of governing coalitions or in exchange for political favours. As such, insulation is a key component of state capacity insofar as it facilitates the formation of complex and long-term state projects.Footnote 10 Among other things, insulation allows for the concentration of technical expertise within state agencies and for the employment of scarce resources towards the pursuit of broader state or societal interests. When bureaucracies belong to patronage networks – that is, when office holders appoint bureaucrats for political reasons – state agencies are more likely to lack expertise and to divert resources towards the pursuit of narrow political objectives.
The potential for bureaucracy-driven reform further increases if state agencies enjoy broad mandates to pursue policy outcomes. By broad mandates I mean the issuing of general orders to address existing problems that do not specify policies or strategies to accomplish them. No state agency is completely able to determine its own goals, but when office holders create broad mandates they free bureaucrats to select personnel, to design alternative policies and policy-making strategies, and even to influence the preferences of those who appointed them.Footnote 11 Such mandates also allow policy elites to pursue linkages with private entities in order to gain resources and lobbying power for their projects.Footnote 12 Broad mandates, in short, transform what could be a top-down relationship between office holders and their bureaucratic agents into a more reciprocal one where influence travels both ways. In contrast, narrow mandates bind agencies to specific policies and policy-making strategies that limit their ability to innovate, to develop ties with other sectors and to have an independent impact on policy.
Bureaucratic autonomy is crucial for sentencing reform because the predicament that demands novel approaches to punishment also discourages politicians from considering non-prison sentences as potential solutions. As Garland argues for Britain, harsh penal laws are both cause and consequence of social instability: they are a response to rising criminal activity, but they trigger the emergence of a penal crisis that imposes heavy financial costs without lowering crime rates or a general sense of fear.Footnote 13 Alternative penalties for certain conducts present a viable solution to this demoralising of state authority, but continuing security concerns induce the public to punish elected officials who propose the weakening of existing sentencing rules. Given these electoral constraints, sentencing reform tends to emerge through the actions of state agencies that work ‘well away from the gaze of the media and political actors’.Footnote 14 Without insulation and broad mandates, however, these agencies are unlikely to promote the unpopular policies analysed here. Rather, we should expect them either to lack sufficient professional expertise or to uphold the electorally constrained preferences of elected officials.
Public opinion in Latin America also lowers the probability of sentencing reform in the absence of autonomous bureaucracies. In 2010 almost 36 per cent of respondents in a Latinobarómetro survey believed that increased punishment is the most effective policy in public security.Footnote 15 In contrast, fewer than 30 per cent believed that social expenditures, civic participation or a stronger judiciary are as productive. Reasons for these attitudes include popular beliefs – exacerbated by victimisation and a media-induced sense of insecurity – that incarceration is a proper punishment, that human rights are ‘privileges for bandits’ and that liberal reforms weaken the police.Footnote 16 Since individuals who equate punishment with effective criminal justice are more politically active, and since politicians are prone to latch on to social fears of crime for electoral reasons, this attitude leads local legislators to uphold tough penal sanctions despite their social and financial costs.Footnote 17 As argued for Argentina, ‘public policy responses to security issues may tend to focus on punitive measures and neglect social aspects because public officials view the problem of insecurity primarily as a political demand that requires a quick response’.Footnote 18
It should be stressed that bureaucratic autonomy does not guarantee sentencing reform. The occasional heightening of media attention to security issues can provoke presidents to issue narrow mandates to the bureaucracy. Also, if other state actors such as judges and prosecutors display a bias against the adoption of alternatives to incarceration, reform becomes less likely. As will be seen below, Brazilian bureaucrats have failed to reduce penalties for street-level drug peddlers for these reasons. Finally, autonomy is irrelevant if a country does not possess a large enough pool of professionals with which to populate the state. Rather, the argument is simply that autonomous policy elites make it more likely that alternative sentencing enters the legislative agenda and garners sufficient support for approval. Autonomy, in other words, is necessary but not sufficient for sentencing reform in countries biased toward incarceration.
This focus on bureaucrats does not discount the causal impact of external pressures. As Keck and Sikkink argue, the adoption of human rights legislation in Latin America often follows the formation of transnational networks of activists with sufficient clout to pressure local governments for policy change.Footnote 19 In the case of criminal justice reforms, the formation of such networks is made possible by the presence in the region of international legal activists and by the recent emergence of domestic non-governmental organisations concerned about prison conditions.Footnote 20 Yet these arguments cannot fully identify the causal process by which international pressures translate into actual policy reform.Footnote 21 Since autonomous bureaucrats engage private actors in order to strengthen their projects, it is arguable that these state actors create the conditions under which external pressures can influence policy outputs in the presence of a strong popular bias towards punitive criminal justice.
The argument developed here also acknowledges the potential impact of ideology on reform. Recently empowered left-leaning administrations in the region have rejected ‘zero-tolerance’ approaches to criminality and undertaken the difficult task of reconciling citizen security with the expansion of universal civil rights.Footnote 22 The electoral victories of the Left, however, did not change the political dynamics addressed above. Rather, popular support for mano dura in the region is such that even ideologically committed regimes have tended to support reactive and coercive rather than preventive approaches to criminal justice.Footnote 23 This suggests that parties sympathetic towards softer sentencing frameworks may be more receptive to bureaucratic suggestions for change, but will fail to alter the status quo if they are forced to limit the autonomy of criminal justice bureaucrats for political reasons.Footnote 24
The argument also holds if we accept Wacquant's claim that contemporary penal policies are elements of a cohesive neoliberal project.Footnote 25 As market-oriented reforms erode the benefits of wage labour, he argues, the state expands and diversifies the reach of its penal arm to control a growing marginalised population. In this process there is a division of labour: politicians pursue harsh but popular penal rules and administrators pursue adaptive policies that do not enjoy popular support.Footnote 26 Arguably, bureaucrats who lack autonomy will pursue politicians’ preferred policies and not develop electorally risky alternatives such as alternative sentencing. However, this article prefers not to conceptualise recent penal laws as elements of a broader neoliberal project. Neoliberalism has contributed to a rise in criminality across Latin America and penal laws in the region have disproportionately affected the poor.Footnote 27 Nonetheless, most of the region has not built the kind of minimal state that Wacquant identifies in the United States.Footnote 28 Rather, as the case of Brazil shows, electoral considerations are key drivers of harsh sentencing rules in the region.
The Political Context of Sentencing Reform in Brazil
Brazil's transition to democracy coincided with the building of a highly punitive system of criminal justice. President Fernando Collor de Mello, the country's first elected executive in over two decades, took office in 1990 and later that year sanctioned a Heinous Crime Law (Lei dos Crimes Hediondos) that prohibits bail, pardons, amnesty and provisional release for kidnappers, rapists, terrorists and drug traffickers. It also prohibits the granting of commutations and the transference to halfway houses for those found guilty of these crimes. Since this law came into effect penal legislation in the country has continued to display a bias towards incarceration. We observe that 47 per cent of all criminal justice norms approved between 1989 and 2006 increased minimum prison sentences, limited the possibility of sentencing commutations or created new categories of criminal behaviour. In contrast, only 23 per cent expanded prisoners’ rights, broadened the scope of alternatives to imprisonment or facilitated sentence commutation.Footnote 29
These trends have persisted despite evidence of their inefficacy.Footnote 30 Between 1995 and 2011 Brazil's incarceration rate increased from 95.5 to 270 prisoners for every 100,000 inhabitants.Footnote 31 This rate imposes heavy financial costs upon the state, especially given the system's overall inefficiency. According to a congressional report, in the mid-2000s the country spent a monthly average of US$ 670 per prisoner compared to the Latin American average of US$ 184.Footnote 32 Prison riots have also been a constant in the current democratic period, notable examples being the 1992 riot in São Paulo's Carandiru prison complex and the 2006 riots that arose as part of a more general wave of gang violence. More recently, prison riots in 2013 in the northern state of Maranhão renewed international attention to the inhuman prison conditions in the country and displayed the government's inability to control prisoners. Violent crime, meanwhile, has not lessened. Homicide rates grew by 20.3 per cent between 1990 and 2000 and have remained consistent since then.Footnote 33
In view of this data, what explains the continuing bias towards incarceration? Brazilian legislators are not immune to zero-tolerance discourses emanating from the United States.Footnote 34 But the country's penal laws cannot be fully understood without an assessment of the public's attitude towards penal legislation. In 1989 and 1990 the local media were saturated with stories about an ongoing crime wave in Brazil's cities that created a strong popular demand for harsh state responses.Footnote 35 In fact, the Heinous Crime Law came into being on the heels of the widely publicised kidnapping of two prominent businessmen. Popular attitudes towards crime and punishment have remained stable since then, and in a 2011 survey 54 per cent of respondents agreed that harsher penalties reduce criminality.Footnote 36 Brazilians also display strong support for life imprisonment and for the lowering of the age of criminal responsibility.Footnote 37
Public attitudes have influenced elected officials, as is suggested by the time lapse between proposal and approval of different types of criminal justice legislation. In September 1989, during the administration of José Sarney, a conservative politician and former member of the military's base of support who presided over Brazil's transition to full electoral democracy, the Brazilian executive drafted an initial Heinous Crime Law bill that spent three months serving as the basis for other projects in the legislature. Once the high-profile kidnappings mentioned above hit the media, however, a similar proposal from the Senate took only 71 days to navigate Congress and move to President Collor's desk for sanction.Footnote 38 This was not unique: between 1989 and 2006 Congress approved 50 per cent of punitive laws in less than one year. By contrast, in similar time frames it approved less than 16 per cent of laws that expanded defendants’ rights.Footnote 39
Public support for harsh sentencing rules affects legislators across the ideological spectrum. Brazil's Congress is consistently fragmented into multiple parties, but there is no clear correlation between ideology and proposals for tougher rules, although right-wing parties tend to propose the harshest projects.Footnote 40 As revealed by a 2006 survey, Brazilian legislators of all stripes agree that the development and pursuit of criminal justice reform is hindered by a lack of political will, and that the issue does not win votes unless a perceived rise in criminality and a media-induced sense of insecurity create popular demands for harsh legislative responses.Footnote 41 In 1984 the outgoing authoritarian regime enacted a reform that emphasised the rehabilitative purpose of penal legislation and established the possibility of non-prison penalties for small crimes, but in today's democratic regime legislators have no motivation to uphold the spirit of that earlier attempt to lower incarceration rates.Footnote 42
Nonetheless, and despite the absence of electoral incentives for reform, Congress has seriously entertained a few projects that contradict the zero-tolerance model. Of these, three proposals stand out for their potential impact on incarceration rates. In November 1998, at the end of Cardoso's first term in office, Congress expanded the possibility of alternative sentencing for small crimes. In 2009 the Lula administration came close to adopting the possibility of legal distinctions between street-level drug dealers and large traffickers with the purpose of limiting prison time for the former. Finally, in the first year of Rousseff's presidency the legislature approved a series of alternatives to the pre-trial confinement of non-violent offenders. What accounts for these legislative innovations?
Transnational Networks and Political Ideology
Arguably, the formation of a transnational network of activists could serve as the primary source of pressure for reform. There is a growing ‘knowledge and policy community’ in the country concerned about high rates of incarceration and their consequences.Footnote 43 A national Catholic body, the Pastoral Carcerária (Prison Ministry), for example, has for years acted as the main venue through which civil society can monitor local prison conditions. Academic institutions and think-tanks such as the Núcleo de Estudos da Violência (Violence Studies Centre, NEV) at the University of São Paulo, in turn, collect valuable data on the socio-economic impact of Brazil's criminal justice system. Many of these organisations have engaged in partnerships with international entities, such as the Ford Foundation and George Soros’ Open Society, which have contributed to the strengthening of these social actors with funding and organisational resources. The synergy between domestic and international activists resembles the kind of boomerang pattern theorised by Keck and Sikkink, whereby domestic actors unable to pressure their governments seek assistance from foreign entities who amplify demands for change and direct them back towards the target country.Footnote 44
Transnational networks of activists can shape domestic debates over prison conditions and contribute valuable knowledge about criminal justice reforms abroad, but they cannot fully explain the changes analysed here. Brazilian civil society did not get involved in criminal justice issues in general, or questions about citizen security in particular, until after the Cardoso administration.Footnote 45 Furthermore, it is unclear how transnational networks can effectively pressure elected officials to change sentencing rules, given the political context discussed above. First, these networks tend to be more successful when dealing with problems that have clear culprits and whose victims can generate some sympathy.Footnote 46 Yet there are many structural reasons for prison overcrowding, and common criminals are not seen as victims in the same way as political prisoners are. Second, since the United States and many other nations also subscribe to zero-tolerance approaches to criminality, countries such as Brazil become less susceptible to calls for them to adopt sentencing reforms as a way to enter a ‘normative community of nations’.Footnote 47
The ideological orientation of Brazilian presidents since 1995 also provides a partial explanation for the reforms observed. President Collor de Mello belonged to a traditional political family who worked alongside the military regime that governed the country from 1964 until the mid-1980s. Cardoso, Lula and Rousseff, on the other hand, began their political careers in opposition to the mano dura policies of the authoritarian regime and their official projects of government included proposals to reconcile citizen security with civil rights protection. Further supporting the importance of the president's ideological orientation is the fact that the 1988 Constitution granted the executive numerous mechanisms with which to dictate the country's legislative agenda, including the right to request urgency for bills in Congress. In Brazil's fragmented party system, this framework improves the odds of legislative approval as long as the executive maintains an ideologically diverse coalition through the distribution of cabinet positions to member parties.Footnote 48 Perhaps unsurprisingly, it was the executive branch that sponsored the proposals for alternative sentencing and limits to pre-trial incarceration.
It is not unrealistic to expect left-of-centre presidents to lend their weight towards the approval of measures designed to solve penal crises that undermine state authority and cause civil rights violations, but an explanation centred on the Brazilian presidency is incomplete for two reasons. First, presidents since Cardoso have built their coalitions primarily to pursue macro-economic stability and limited market reforms that enjoy a relatively broad political appeal. Criminal justice, in contrast, belongs to a list of residual and controversial policy areas that enjoy marginal partisan support and thus receive only partial executive attention.Footnote 49 Second, the executive is also subject to the pressures for tough measures against crime, as indicated by the fact that 50 per cent of laws raising penal sanctions between 1988 and 2006 originated in the executive and not in Congress.Footnote 50 The Brazilian presidency may support new sentencing rules, but we still need to identify the causal links in the process by which reforms overcome existing political and electoral barriers. To that end, the next section identifies Brazil's bureaucratic agents of reform, measures their levels of autonomy during the three presidential administrations mentioned above, and traces the process by which sentencing reform navigated Brazil's political system. Since Rousseff was Lula's hand-picked successor, the article discusses the two Workers’ Party administrations simultaneously.
Bureaucratic Autonomy and Sentencing Reform in Brazil
Bureaucrats in a position to push criminal justice reforms are numerous in Brazil's Ministério da Justiça (Ministry of Justice, MJ). Three agencies, in particular, shape and drive the ministry's policy pursuits. The Secretaria de Assuntos Legislativos (Secretariat of Legislative Affairs, SAL) is responsible for drafting and monitoring legislative projects relevant to the ministry. The Secretaria Nacional de Justiça (National Secretariat of Justice, SNJ), in turn, is responsible for coordinating justice policy with other state institutions, the legislative and the judiciary, and local governments. Finally, the head of the cabinet – in addition to having the ability to either undermine or boost the actions of the SAL and the SNJ – also appoints members to the Conselho Nacional de Política Criminal e Penitenciária (National Council on Criminal and Penal Policy, CNPCP). The CNPCP is responsible for developing guidelines on the serving of sentences and on identifying priorities for criminal and penal policy.
The Brazilian presidency appoints all three positions.Footnote 51 Thus, whether the SAL, the SNJ and the justice minister enjoy autonomy depends on whether the president insulates them from patronage networks and grants them broad mandates to shape policy. Because insulation means that the president does not appoint individuals solely in return for political support, we should expect insulated bureaucrats to be either career bureaucrats or members of the president's own party (although the executive might appoint individuals from coalition partners who possess some degree of technical capacity). Political science has yet to develop an objective way to measure bureaucratic mandates.Footnote 52 Nonetheless, I argue that broad mandates should be associated with interactions between the bureaucracy and private actors in the development of policy responses to existing problems, and with occasional conflicts between the bureaucracy and the president.Footnote 53 A micromanaged bureaucracy lacks the freedom to reach outside the state for policy proposals and is precluded from developing projects that conflict with office holders’ political and electoral interests.
Bureaucratic autonomy in the Cardoso administration
When filling bureaucratic positions, elected officials must balance the political needs of governance with the technical requirements of policy-making.Footnote 54 We observe this dilemma at play during President Cardoso's two terms in office (1995–8 and 1999–2002). Cardoso relied more heavily on technical criteria than his three civilian predecessors when appointing cabinet members, but he could not completely ignore the political realities of Brazil's multiparty system. Engaged as it was in the pursuit of macro-economic stability and market reforms, his administration insulated the finance ministry but allowed other agencies – including the Ministry of Justice – to serve as currency in the management of governing allies in Congress.Footnote 55 The periodic reshaping of coalition partners accounts for Brazil having ten justice ministers in the eight years of Cardoso's presidency, excluding interim ministers.
Not all of Cardoso's cabinet appointments to the MJ served a purely political purpose, however. Occasionally, he selected individuals who could facilitate coalition-building but also possessed the technical capacity usually associated with insulated bureaucratic agencies.Footnote 56 Cardoso's first and longest-serving justice minister, Nelson Jobim (January 1995–April 1997), belonged to this group. He was a key member of the coalition's main party and had the political background necessary to manage the relationship between the president and his allies.Footnote 57 Yet Jobim was also a lawyer and former congressman who held a leadership position in Brazil's constitutional assembly of 1987–8 and who directed a project of constitutional revision in 1993 and 1994. His two successors during Cardoso's first term lacked such credentials, but two ministers in the president's second administration were also legal professionals with a commitment to reform.Footnote 58
Bureaucratic insulation waxed and waned at the top of the ministry, but it remained constant in the second rung. There are no systematic data on the composition of federal agencies during Cardoso's tenure, but only two individuals led the SNJ between 1995 and early 2002, and both were career bureaucrats or academics without close connections to political parties. Sandra Valle, a career bureaucrat in the Congress, took the position under Jobim and held it until the first year of Cardoso's second term. Her replacement, Elizabeth Sussekind, is a criminologist with experience in academia and in the criminal justice bureaucracy of Rio de Janeiro. Another bureaucrat held the position for the last few months of Cardoso's presidency.
The maintenance of a strong technical presence within the SNJ meant that the occasional appointment of justice ministers for political reasons could hinder, but not completely obstruct, the bureaucratic development of policy responses to the country's penal crisis. Specifically, insulation at the second level kept experts in a position to influence the preferences and shape the actions of the justice minister. As stated by a former head of the SNJ, the secretariat would develop proposals, negotiate support from relevant actors across Brazil, and withhold their presentation to the head of the cabinet or the president's staff until bureaucrats could effectively ‘sell’ the projects as being necessary given existing conditions in the country, and as having the potential to improve the minister's political standing.Footnote 59
Insulation from political patronage networks does not facilitate bureaucracy-driven reform if the president micromanages state agencies, but the justice ministry's interactions with foreign actors during Cardoso's first term suggest that bureaucrats had a free hand in developing and advocating policy responses to Brazil's penal crisis. In early 1995, for example, Jobim and ministry bureaucrats attended the ninth United Nations Crime Congress in Cairo to collect information on (among other things) successful strategies for dealing with organised crime. Based on the knowledge acquired, the ministry developed legislative projects to improve the detection and prevention of money-laundering, to create a legal basis for wire-tapping and to strengthen plea-bargaining practices.Footnote 60 Civil society during the Cardoso administration was not sufficiently involved in public security issues, but that did not prevent the MJ from engaging other actors for support.Footnote 61
Occasional clashes between the ministry and the executive also indicate the absence of political micromanagement. During Cardoso's second term, for example, the minister, José Carlos Dias, proposed the repeal of the Heinous Crime Law, the creation of specialised drug courts and a general rejection of incarceration as the default penalty for rule breaking. This minister had no ties to political parties, had worked professionally in the defence of human rights, and gave carte blanche to groups within the ministry to overhaul the country's penal system.Footnote 62 But, unlike Jobim, who promoted penal reforms piecemeal to avoid congressional backlash, Dias attempted to induce the legislature to adopt a complete overhaul of the system. The Cardoso administration quickly rejected the argument that these proposals reflected the government's preferences, and ousted Dias from the cabinet as soon as conservative members of the coalition expressed opposition to his project.Footnote 63 The fact that these controversial proposals made it to both the media and the legislature despite the opposition that they were sure to attract strongly suggests that bureaucrats enjoyed broad mandates to develop and promote them.
Bureaucrats and alternative sentencing
Brazil first adopted alternatives to incarceration for small crimes during the country's slow transition to democracy. In 1984 the Ministry of Justice and human rights advocates took advantage of the weakening of conservatives in the military regime to establish the possibility of non-prison penalties for crimes carrying less than one-year sentences.Footnote 64 In 1996 Jobim's Ministry of Justice developed a bill that expanded the reach of alternative sentencing to include first-time offenders whose penalties amounted to less than four years’ imprisonment. Cardoso signed the alternative sentencing bill into law as Law 9714/98 in November 1998, more than two years after its original drafting.
The reality of Brazilian prisons justified the project. Census data show that in 1995 the country had a deficit of 75,887 places in the prison system.Footnote 65 More importantly, 18 per cent of prisoners had received sentences of one to four years. As another study shows, in 1996 about a quarter of the prison population in the state of Rio de Janeiro had not committed a violent crime.Footnote 66 Because their sentences carried more than one-year terms, however, they fell outside of the scope of the 1984 law. As the justice minister who had sponsored that earlier law stated, the 1996 project adopted the same reasoning they had used in 1984: namely, that because of prison overcrowding and the problems it caused, the state should limit incarceration to the punishment of serious or violent criminals.Footnote 67
Bureaucratic autonomy facilitated the genesis and approval of Law 9714 for two reasons. First, it allowed for the concentration of expertise in the agency. In particular, Jobim transformed the National Council on Criminal and Penal Policy into a strong agent of change. In 1990 the CNPCP had argued in favour of harsher penalties for serious crimes as a proper response to popular clamours for security,Footnote 68 but after Jobim altered its composition the council developed the blueprint for the first attempt to mitigate the impact of the Heinous Crime Law upon prison conditions.Footnote 69 The appointment of individuals with experience inside the Brazilian penal system, including a director of Brazil's Department of Corrections and a former director of Rio de Janeiro's correctional system, triggered this change of attitude. Retired legal professionals with little or no experience in prison matters dominated the council prior to Cardoso's presidency, but Nelson Jobim strategically selected individuals who could credibly make the case for changes to existing sentencing rules.Footnote 70
The CNPCP advocated for the bill in a number of ways. Between 1996 and 1997 its members published a number of academic and news articles defending the idea that prison terms should be set aside for violent criminals. In addition to the overcrowding problem, these articles brought attention to the lower rates of recidivism among individuals sentenced to alternative penalties and the high economic cost of incarceration.Footnote 71 The CNPCP also sponsored public meetings with foreign policy-makers who shared their experiences regarding alternative sentencing laws, and worked with the German Konrad Adenauer Foundation to release an edited volume on the issue.Footnote 72 Finally, the council's president sponsored an unprecedented census of the prison population, which revealed the demographic data mentioned above. The study concluded that overcrowding was the penal system's most severe problem, that the creation of more prisons would only serve as a stopgap, and that alternative sentences for non-violent criminals could reverse incarceration rates.Footnote 73 The CNPCP released the census results to Brazil's most widely circulated weekly magazine, which acknowledged the data in an article that refers to Brazil's prison situation as ‘catastrophic’, calls penitentiaries ‘beast factories’ and compares the government's alternative sentencing project to similar and positive experiences abroad.Footnote 74
Second, the autonomy of the National Secretariat of Justice prevented the project from being undermined by changes at the cabinet level. Nelson Jobim provided a strong defence for the alternative sentencing bill in the Brazilian media, but he left the ministry before legislative approval was obtained.Footnote 75 His successors during Cardoso's first term served primarily to strengthen the ruling coalition and not to boost the technical aspects of the ministry. These individuals had no real vested interest in the reforms in progress, but the politically insulated SNJ enjoyed a sufficiently broad mandate to continue developing support for the alternative sentencing law among different interest groups across Brazil. The SNJ also interacted with the new ministers, with Congress and with the executive as the project underwent multiple reviews.Footnote 76 In 1998, for example, the SNJ organised a meeting between the minister, Renan Calheiros – from a centre-right party in Cardoso's coalition – and a former member of the CNPCP. After the minister's initial display of indifference towards alternative sentencing, the council member opened a box of disposable diapers and proceeded to explain how the theft of that product had landed a woman in jail for two years and cost taxpayers over seven thousand dollars.Footnote 77 The minister relied on similar arguments to defend the law when Congress approved it in November.Footnote 78
According to Human Rights Watch, the publicity surrounding the many prison riots that took place in the early 1990s triggered the bill's approval.Footnote 79 But, from the theoretical perspective laid out here, the publicity around the riots created the opening for a bureaucracy-sponsored project to circumvent the political barriers to sentencing reform – ironically, since the Brazilian media are generally seen as contributing to a general sentiment of fear. As noted above, Brazilian office holders have few electoral motivations to invest in the development of alternatives and limits to incarceration. Evidence of this is the fact that the presidency requested urgency for the bill only four weeks before the presidential election of 1998 – meaning that it would come up for a vote and for presidential sanction after the ballots had been counted – and the fact that party leaders in the House of Representatives agreed to keep individual votes on the measure secret when it first came up for deliberation in 1997.Footnote 80 If bureaucratic agents of change had not laid the groundwork for the new law and used existing crises to strengthen their arguments, it is less likely that the prison riots would have triggered such changes to the country's sentencing structure.
Bureaucratic autonomy under Lula and Rousseff
Presidents Lula and Rousseff of the Workers’ Party belong to Latin America's moderate Left, which seeks to promote social equity within the constraints of the market economy.Footnote 81 Their administrations maintained the general contours of Cardoso's macro-economic policies, but unlike their predecessor they did not fill cabinet positions according to the requirements of coalition-building. Rather, the PT has consistently controlled more ministries than coalition parties with a larger congressional representation.Footnote 82 This approach to cabinet formation insulated many agencies from patronage networks, and as a result only three justice ministers have overseen Brazil's criminal justice system since 2002. Márcio Thomaz Bastos, who served during Lula's first term (2003–6), is a criminal defence lawyer with no partisan affiliation. Tarso Genro of the Workers’ Party, minister during Lula's second term (2007–10), had previously acted as party president, as mayor of a major Brazilian city and as minister of education. Finally, Rousseff's current minister, José Eduardo Cardozo, is a former PT congressman with a history of involvement in judicial and criminal justice reform projects.
Recent data on the composition of Brazil's state agencies indicates that Lula and Rousseff have insulated the second tier of the federal bureaucracy as well. Over 70 per cent of high-level bureaucrats appointed during the two Lula terms – the so-called DAS 5 and 6 officers – were career bureaucrats and only 25 per cent of the total had a party affiliation. Of these, four-fifths belonged to the Workers’ Party.Footnote 83 In the Ministry of Justice, data show that during Lula's second term almost 85 per cent of appointees to the upper levels of the bureaucracy lacked a party affiliation, and most of those who did have one belonged to the PT.Footnote 84 These numbers indicate that appointees to the ministry have tended to be policy experts with a bureaucratic career, and that their selection is not merely used for patronage purposes.Footnote 85 This conclusion also applies to Rousseff's administration: as a recent attempt to measure state capacity in the country shows, in 2011 the Ministry of Justice could be considered one of Brazil's ‘islands of bureaucratic excellence’ on account of the degree of expertise of its members, their lack of partisan connections and their average longevity in the civil service.Footnote 86
There are indications that this politically insulated federal agency has also enjoyed broad mandates since 2002. Lula, for example, dismissed a team of scholars and bureaucrats whose public security project, aimed at boosting the federal government's involvement with policing issues in the states, raised the electoral risk for the presidency in case of policy failure.Footnote 87 As detailed below, Rousseff removed a top-level bureaucrat following his open support for the expansion of alternative sentencing to include some drug-related crimes. Arguably, if the Ministry of Justice lacked broad mandates we would not have observed these instances of conflict. Also, as explained below, the ministry has actively engaged with non-state actors in the development and promotion of criminal justice projects since Lula became president.
Bureaucrats and drug-related crimes
In August 2006 Congress approved Brazil's current drug law (Law 11343/06). One of its central features is the elimination of incarceration for drug users, including repeat offenders; these are now subject to community service and educational measures. However, hard-line legislators conditioned their support for this measure on the establishment of increased penalties for dealers.Footnote 88 Specifically, this law raised the minimum prison sentence for drug dealing from three to five years. First-time offenders not associated with organised crime can have their jail sentences reduced to less than four years, but, because the law does not enunciate objective distinctions between small dealers and members of trafficking organisations, it is easy for judges to reject these reductions. Furthermore, even if judges grant reductions the law prohibits the substitution of alternative sentences for prison time. Finally, the law does not provide clear criteria for distinguishing users from dealers, thus opening the door for the continuing incarceration of drug addicts.
Just as in much of Latin America, harsh penalties for drug-related crimes and the absence of objective distinctions between low- and high-level traffickers have triggered a demographic explosion inside Brazil's penitentiaries.Footnote 89 In 1995 slightly over 11 per cent of the country's prison population was serving time for drug trafficking.Footnote 90 In 2005 this percentage was only one point higher, but by the end of 2011 drug dealers accounted for almost a quarter of all prisoners.Footnote 91 Because most of those arrested are small dealers and drug ‘mules’ of low socio-economic status, the 2006 law has also exacerbated the historical tendency to criminalise poverty and to accentuate existing socio-economic inequalities in the country.Footnote 92 For example, individuals of a low socio-economic status are less likely to be characterised as users than middle-class individuals caught with similarly small quantities of drugs.Footnote 93
Political support for sentencing reform concerning drug-related crimes is weak because Brazilian society demonises participants in the drug trade and demands harsh treatments for them.Footnote 94 Cardoso, for example, vetoed a 2002 version of the drug law on the grounds that some of its language could potentially cause the release of individuals serving time for drug-related crimes.Footnote 95 It is worth noting, however, that since leaving office Cardoso has joined other former Latin American heads of state in an Initiative on Drugs and Democracy that openly advocates the decriminalisation of drug use and the creation of clear distinctions between small drug dealers and violent criminals or large-scale traffickers.
Lula's appointment of Márcio Thomaz Bastos to the MJ motivated the development of bureaucratic efforts to alter the existing sentencing framework for drug dealers. Bastos appointed CNPCP members who consistently opposed proposals to increase prison penalties and who drafted wide-ranging Christmas pardons for small dealers who had already served some time.Footnote 96 The Secretariat of Legislative Affairs, in turn, attempted a two-step strategy to circumvent barriers to the implementation of legal distinctions between small and large dealers. First, it pursued connections with academic institutions to develop normative as well as functionalist arguments in favour of sentencing reform. Early in the second Lula administration, for example, the SAL partnered with the United Nations Development Programme (UNDP) to build formal ties between the ministry and academia. The ensuing Projeto Pensando o Direito (Thinking about Law Project) provides public funding for empirical and theoretical research on policy areas deemed critical by the ministry's bureaucracy. In 2008, SAL-sponsored research analysing international doctrines on the proper penal treatment for small dealers assessed how the judiciary applied the sentencing regulations of the drug law, and proposed reforms to the existing policy framework. The study laid out the socio-economic costs resulting from the lack of clear distinctions between small and large traffickers and recommended, among other things, the possibility for alternative sentencing and sentence commutation for non-violent and first-time offenders.Footnote 97 According to the study's chief investigator, bureaucrats in the Ministry of Justice and legislators with a history of support for non-draconian legislation worked behind the scenes to allocate funding for academic research on the drug law.Footnote 98 The ministry then presented the study to Congress and in multiple conferences.
Second, the SAL manoeuvred to attach the possibility of alternative sentences for street-level dealers to a bill that raised the minimum amount of time served for certain crimes before prison sentences could be converted. Supporters of clear distinctions between small and large traffickers expected this legislative ‘piggyback ride’ to place them in a better position for bargaining with conservative members of Congress on changes to the status quo.Footnote 99 The strategy worked at the outset, and the SAL successfully brokered a deal by which they would support the requirement that parole recipients wear electronic monitors in return for new rules about sentences for small traffickers.Footnote 100 However, notwithstanding these advances, in November 2009 a senate commission rejected alternative sentences for small drug dealers. The decision to avoid open confrontation with supporters of the status quo coincided with the proximity of an election year and a rise in the potential electoral cost for Lula's party of open support for softer penalties for heinous crimes in general, and for drug dealers in particular.
The election of Lula's hand-picked successor in 2010 allowed for some bureaucratic continuity and renewed efforts for sentencing reform. Rousseff's appointment to the Ministry of Justice, former representative José Eduardo Cardozo, maintained a number of individuals from the previous administration inside the ministry's agencies. More crucially, he singled out a former head of the SAL to lead the Secretaria Nacional de Políticas sobre Drogas (National Secretariat of Drug Policy, SENAD). Soon after his appointment the new SENAD chief gave a candid interview in which he defended alternative penalties for small traffickers.
The president's reaction to this interview points to the often conflicting relationship between electorally constrained office holders and autonomous bureaucrats. Rousseff's administration immediately pushed for the secretary's resignation and effectively closed the door for sentencing reforms towards drug-related crimes. However, recent developments suggest a renewed impetus in the Ministry of Justice for policies designed to limit the incarceration of street-level, non-violent drug dealers. In April 2013 the minister sponsored the appointment to SENAD of a public defendant who has expressed ‘enthusiasm’ towards alternative sentencing and support for sentencing reductions for street-level dealers.Footnote 101 Soon afterwards the MJ pressured members of Congress to change a legislative project that would further increase the minimum prison sentence for drug trafficking to eight years. Specifically, the ministry has asked Congress to develop clear and objective distinctions between street-level dealers and large traffickers as a way to exclude small dealers from this increased penalty. Meanwhile, the legislator who drafted this harsh bill has openly accused the ‘second tier of the justice ministry’ of defending the decriminalisation of marijuana and of undermining attempts to continue the war on drugs.Footnote 102 Bureaucratic reformers have failed to promote softer sentencing rules for drug-related crimes, but they remain the main agents of reform and the actors most likely to hinder the legislature's more draconian tendencies.
Bureaucrats and pre-trial incarceration
President Rousseff's response to SENAD's position on small dealers displayed anti-reformist undertones, but within four months of her inauguration Congress approved Law 12403/11. This law creates alternatives to the pre-trial imprisonment of those accused of non-violent crimes, and thus can have a positive impact on Brazil's penal system. Pre-trial incarceration tends to target the poor and marginalised, push families towards poverty, and impose heavy fiscal costs upon society.Footnote 103 The sheer number of individuals involved magnifies these consequences. By the end of 2010, over 164,000 of Brazil's 445,705 inmates, or 37 per cent of the total prison population, awaited trial. This percentage had remained virtually unchanged since 2004.Footnote 104 More troubling, in 2010 the system had a deficit of 164,624 beds.
Despite these statistics, legislators excluded alternatives to pre-trial incarceration from a basket of criminal justice reforms approved during Lula's second term. The Cardoso administration had originally sent a series of changes to Brazil's codes of civil and criminal procedure to Congress in 2001, but they only resurfaced as part of a broad agreement between the three branches of government to promote efficiency-enhancing judicial reforms. Given Brazil's ongoing transition to a market economy, the first wave of changes modified the country's civil code according to suggestions by the World Bank.Footnote 105 The government only prioritised criminal procedure reforms in 2007 – after Lula's successful re-election bid – and only after the highly publicised death of a young boy during a car-jacking incident in Rio de Janeiro. Media coverage of this tragedy triggered proposals for a lower legal age of criminal responsibility, but the MJ took advantage of this newly rekindled attention to penal issues to bring the pre-existing reform projects up for discussion.Footnote 106 Multiple political barriers for reform, however, had to be addressed piecemeal by the government. Given its controversial nature, the proposal for the creation of alternatives to pre-trial incarceration fell to the back of the queue.Footnote 107
Numerous factors allowed bureaucratic actors to get the bill deliberated and approved quickly after Rousseff's inauguration.Footnote 108 The project was not a priority for the government but justice minister José Eduardo Cardozo had presided over the congressional commission that assessed the project's constitutionality during Lula's second term. He maintained constant dialogue with the SAL during that period and once in the cabinet gave political support for the secretariat's policy pursuits. These pursuits, in turn, benefited from the fact that congressional leaders of the Workers’ Party had appointed sympathetic legislators to the congressional committee on public safety, which was normally populated by former prosecutors and police investigators who tended to support draconian policies. Finally, reformers took advantage of the fact that other aspects of the law drew legislators’ attention away from the provision on provisional incarceration. In particular, legislators centred their opposition on the proposal to end so-called ‘special prisons’ for inmates with college degrees. Policy reformers used the provision as a bargaining chip and conditioned its removal with the maintenance of alternatives to provisional incarceration.
Civil society played an important role in this process. Three organisations, in particular, developed a monitoring and lobbying project that compounded the SAL's reformist strategy. The Sou da Paz institute based in São Paulo first emerged in the late 1990s as a social movement in favour of strict gun control laws. In 2008 George Soros’ Open Society contacted the institute to propose the formation of a network of organisations in support of sentencing reform.Footnote 109 The network included Conectas, a human rights organisation, and the Prison Ministry already mentioned. Once formed, this network developed a database of existing legislative proposals in order to identify bills that fit their overall reformist goals. To that end they contacted the SAL, which proposed assistance in return for empirical data on criminal justice issues of interest to the ministry. The resulting partnership pressures Congress in four different ways: it identifies rapporteurs for congressional projects of interest, it lobbies for the timely release of legislative reports on existing projects, it presents empirical data on the connections between existing policies and objective prison conditions, and it monitors sudden changes to proposals in the legislative pipeline.Footnote 110 Because of its advanced legislative status, the provisional incarceration bill became the network's first target and in January 2011 Sou da Paz hired a lobbyist to help perform these monitoring and pressuring functions. Prior to the law's approval, and partially to pressure lawmakers, the coalition carried out a number of public actions to raise awareness about the plight of non-convicted individuals inside Brazil's prisons.Footnote 111
The process behind the law's enactment indicates the evolution of a strong transnational structure of support for criminal justice reform in Brazil. This development is partially due to the growing interest in controversies about criminal justice among academic and civil society circles in the country and abroad and their active attempts to influence the legislative process on the matter. The success of this coalition in pushing an unpopular piece of legislation through political gridlocks, however, hinged on the autonomy of the state agencies involved. If office holders had micromanaged these agencies, or if Lula and Rousseff had politicised the Ministry of Justice, it is less likely that Congress would have debated and approved this reform. Civil society may use this learning process in the future to push for changes even in the absence of an autonomous and supportive bureaucracy, but current concerted efforts to change the system remain very much centred on the state agencies that monitor the overall system and can exercise some level of influence over elected officials.
Conclusion
As Guillermo O'Donnell emphasised many years ago, Latin American democracies combine well-protected political rights with inchoate civil rights.Footnote 112 That is, democratic participation is widely respected but the liberal component of democracy is undermined by the fact that not all citizens enjoy equal, fair or proper treatment from the state. Harsh sentencing rules contribute to this democratic deficit: they fuel the humanitarian crisis inside Latin America's overcrowded prisons and disproportionately affect less privileged members of society. Unfortunately, democracy itself helps explain this outcome. Specifically, a widespread sense of insecurity creates popular pressures for elected officials to adopt zero-tolerance approaches to crime. Because those most affected by harsh rules are unable to influence the political process, these pressures result in criminal justice systems biased towards highly punitive penal policies.
This article describes the causal process by which attempts to alleviate demographic pressures upon Brazil's prison system emerged and navigated a political environment biased towards severe punishments. The analysis traces reform projects to the actions of autonomous policy elites who rely on their position in the state to overcome political barriers to the adoption of softer sentencing rules. These elites may act out of a sincere concern for the well-being of marginalised populations, but it is also clear that they act away from the public eye to solve crises that threaten the authority, stability and financial health of the state itself. Whatever the case, the analysis of sentencing reform in Brazil supports the argument that it is through the actions of autonomous state agencies that attempts to benefit the underprivileged in Latin America – prisoners included – can overcome the short-time horizons of politicians and the lack of solidarity among the electorate.Footnote 113
Political science has tended to overlook the possibility of bureaucratic policy-making in Latin America because it is assumed that local policy elites are merely the agents of elected officials.Footnote 114 But, despite the region's historical record of excessive bureaucratic subordination, the analysis indicates that state autonomy should not be discounted as a potential explanation for policy reforms in general, and criminal justice reform in particular. Rather, the actions of Brazil's Ministry of Justice in developing proposals, pursuing connections with private actors, and acting strategically to improve reforms’ chances of success indicate that the relationship between bureaucratic agents and their political principals can be much more reciprocal than is usually expected, and that influence can also flow from the state to elected officials.
What accounts for the observed autonomy? It is not the purpose of this article to explain office holders’ decision to increase bureaucratic insulation and issue broad mandates, but it is worth exploring potential explanations for our anticipation that this will be the case in criminal justice matters. Poor performance in combating violence is a primary reason for the relatively low levels of support for democratic governments in Latin America.Footnote 115 Given the electoral backlash that failed criminal justice policies and continuing perceptions of insecurity entail, electoral winners have incentives to issue broad directives to state agencies that work on these matters as a way to avoid blame. The fact that bureaucratic autonomy was relatively higher during the two PT administrations – as indicated by the stronger insulation of the justice minister himself – supports this assertion. Presidential candidates have avoided debates on criminal justice issues because governors are responsible for much of public security policy in Brazil; yet during Lula's 2006 presidential bid criminal justice became a topic of debate for the first time, as a result of gang-orchestrated violence in São Paulo.Footnote 116 Presidents may still limit autonomy for political or electoral reasons, but as long as the requirements of coalition-building and the proximity of elections do not oblige them to do so we should expect bureaucratic autonomy in the criminal justice sphere to thrive.
The argument that office holders have incentives to expand the autonomy of the criminal justice arm of the state opens interesting avenues for future research. This article focuses on policy elites at the national level, but much criminal justice reform takes place away from the centre of power. The process by which certain types of deviance get lower penalties, for instance, often happens at ‘the “shallow” and hence less visible end of criminal justice’.Footnote 117 Thus, it would be worthwhile to explore whether and how prison administrators, public defendants, prosecutors and the police interact at the subnational level to either reinforce or assuage the impact of existing punitive legislation. It would also be worthwhile to investigate whether and how some of these agencies interact with private actors to monitor the implementation of more recent legislative innovations.Footnote 118 As the crisis of Latin American criminal justice deepens, it is not unrealistic to assume that dense and reciprocal connections will appear between state agencies and civil society in support of changes to existing punitive frameworks.
