Beato, C and Silveira, A M 2014 Effectiveness and Evaluation of Crime stability Prevention Programs in . Stability: International Journal of Security & Development, 3(1): 20, pp. 1-18, DOI: http://dx.doi. org/10.5334/sta.dr

RESEARCH ARTICLE Effectiveness and Evaluation of Crime Prevention Programs in Minas Gerais Claudio Beato* and Andréa Maria Silveira†

This essay analyzes previous studies evaluating the effectiveness of the crime prevention policies adopted by the Government of Minas Gerais (). In this work, greater emphasis is placed on studies evaluating outcomes than on studies dealing with the process of setting up and implementing programs and projects. In order to allow a more systematic discussion, the Maryland Scale, which cat- egorizes research and evaluations according to the methodological strengths and weaknesses in five levels, is employed. Subsequently, the authors draw a parallel between Brazil and other settings. Finally, this essay lays out the implications of this discussion regarding the prevention programs.

Introduction broader reach in the execution of programs A mix of social policy and crime prevention targeting more specific groups such as the programs is often proposed to keep crime young, high-risk families, youths on proba- rates low. Although the association between tion schemes or street dwellers just to name general social programs and specific pro- a few, that can be more directly linked to the grams to fight crime is unclear, this type of issue of violence prevention (Banco Mundial social intervention is commonly associated 2003; Soares 2006). Nevertheless, the with a reduction in crime (Welsh and Far- power of the municipal is restricted rington 2012; Rolim 2006). to , which represents a What are the most important effects of limiting factor for the structuring of ‘Weed these policies for reducing crime? The dif- and Seed’ projects that involve elements of ficulty of evaluating these policies partly qualified enforcement and crime prevention stems from the fact that social policies are through social development (Banco Mundial devised and implemented at different levels 2003; Soares 2006; Rigakos 1996; Kelly et al of public administration (municipal, state, 2005). Consequently, many programs of this and federal). In Brazil for instance, such poli- nature aimed at bringing together different cies are under the control of state govern- control strategies are derived from state-level ments. Municipal districts however, have a agencies, such as UPPs (Pacification Police Units), Pacto pela Vida (Pact for Life) or Fica Vivo (Stay Alive). * Professor of Sociology, Universidade Federal de To make things worse, there is often a prob- Minas Gerais, Brazil lem with how crime prevention programs are [email protected] conceptualized. Treating crime prevention † Professor of Public Health, Universidade Federal de Minas Gerais, Brazil programs distinctly and separately from pro- [email protected] grams connected with policing is a common Art. 20, page 2 of 18 Beato and Silveira: Effectiveness and Evaluation of Crime Prevention Programs in Minas Gerais mistake. Many of the assessment projects focus of such projects greatly reduces their are run by the police or feature the police potential for generalization toward applica- in an active role. The IGESP (Integration and tion in broader policies. Management of Public Security), and GEPAR The Brazilian tradition entails a broader (Group Specialized in Policing Risk Areas) of strategy, with groups of projects comprising a the Fica Vivo (Stay Alive) are highly successful state policy, as opposed to specific programs examples of prevention programs involving within localized institutional settings (Beato the police. The Patrulhamento Ativo (Active 2012; Soares 2006). As a result, in Brazil, Patrol) program has also shown promise, many discussions are centered on evaluat- although more systematic assessment is nec- ing how programs were run rather than their essary (Beato 2008; Soares et al 2013). outcomes. While it is true that this type of Another aspect to consider is the scope of strategy is virtually inevitable, given Brazil´s the interventions, which directly influences context of short-term urgency, institutional the difficulty of conducting evaluations. In and constitutional constraints, and cognitive Latin American literature, particularly from capacity prevailing in public security, it also Brazil, there is a predominance of analyses severely limits the ability to critically assess concerning institutional projects of a broader which components and elements work in and more systemic nature (Beato 2012). crime prevention (Beato 2012; Silveira 2007). Thus, as a general rule, these analyses range This renders the programs more vulnerable from interventions of a managerial nature to the development of inefficient, expensive for integrating actions at different levels, to and largely ineffective policies. There are specific programs and projects aimed at dif- exceptions however, some of which will be ferent audiences. Examples include urban discussed below. restructuring and requalification projects, or schemes for the reeducation and reinte- Prevention of crime and violence in gration of juvenile offenders or individuals Brazil leaving the prison system. It therefore fol- The Federal Government of Brazil began lows that the most renowned policies of this incorporating ‘elements of prevention’ pro- nature are endorsed by state governments, posing interventions upon known causes of which often seek federal funding for them. crime and violence by setting up the Public One of the texts most central to the dis- Security Fund established by Minister Jose cussion of evidence-based crime-prevention Gregorio. More recently, since 2003, the Fed- policies is the work of Lawrence Sherman eral Government has proposed a National (Sherman et al 2002), which draws on over Policy for Public Security whose agenda fea- 500 studies, assessing their effectiveness in tures the prevention of crime and violence. the United States and England. These stud- In particular, the reformulation of the Law of ies are grouped into seven institutional set- the National Fund for Public Security opened tings for evaluation: a) family; b) schools; the door for possible municipal financing c) communities; d) job market; e) places; f) for crime prevention programs. Although police agency; and g) courts and correctional these programs have not yet been formally facilities. The North American and English assessed by the Ministry of Justice, a database cases comprise numerous experiences, pro- has been built to compile, systematize and jects and programs specifically focusing on disseminate the public’s experience with the particular violence and crime-related issues. programs. In 2005, through the document These consist of highly specific projects with entitled ‘Profile of the Practices Recorded at a well-defined focus, enabling robust scien- the Democratic Monitoring Unit of Practices tific evaluation in which causal relationships for the Prevention of Violence and Crime in can be established. However, the narrow Brazil,’ the Ministry of Justice reported the Beato and Silveira: Effectiveness and Evaluation of Crime Prevention Art. 20, page 3 of 18 Programs in Minas Gerais basic characteristics of 168 prevention prac- cent of public respondents deemed the tices registered at the Unit up to May 2005 actions successful (MJ 2005). (MJ 2005). In all, only half of the programs Although the database listed only 168 were found to have undergone some type of initiatives, professionals from the field have evaluation (Silveira 2007). acknowledged the existence of a greater The report revealed that the number of prevention programs and pro- and Fire Service were the institutions that jects in place, many conducted by NGOs, reli- most promoted prevention of violence and gious institutions, and OSCIPs (Civil Society crime, accounting for 56 per cent of all such Organizations of Public Interest). The major- actions. The conducted 23 per ity of these initiatives operate at the commu- cent, while the state secretariats and other nity level with a target population of children NGOs accounted for 11 and 10 per cent of and young adults. Most of the projects offer actions, respectively. The most commonly workshops for recreation, sports, vocational cited problems, particularly by the police training, remedial tutoring, culture and art. forces, were drug trafficking and use along The idea of keeping youths occupied in a with drug addiction. Among state secretari- venue for socializing—and therefore off the ats, Civil Police and other organizations, a streets—is prevalent. third of the projects were dedicated to the In the field of health policy, Administrative prevention of juvenile crime and to promot- Rule no. 737/GM 16 of May 2001 concern- ing the Child and Adolescent Statute (ECA). ing the National Policy for Reducing Morbi- The main obstacles cited in attaining pro- mortality by Accidents and Violent Acts gram objectives included a lack of coordina- plays an important role in establishing dif- tion of ‘counter-advertising’, skepticism of ferent institutional responsibilities, and also the community, interaction problems, lack ‘embrace[s] measures inherent to the promo- of engagement, interest, integration, partici- tion of health and prevention of these events, pation, and reluctance and resistance from through the implementation of processes of the community. Also, some two-thirds of the coordination with different social segments programs reported a lack of human, physical (Brasil 2001).’ This policy is grounded in and material resources as an obstacle to pro- the Brazilian legal concept of health, which gress (MJ 2005). encompasses issues related to lifestyle and to The majority of the programs were devised social, historical and environmental condi- with urban populations in mind. The most tions existing in Brazilian society. common level of intervention was at the com- To this end, the policy’s first directive is pro- munity level, where 12–24 year-olds were the moting the adoption of behaviors and safe target population in 42.9 per cent of interven- and healthy environments, via ‘broad mobi- tions while 55.4 per cent of initiatives involved lization of society at large, and particularly victims and aggressors. Some 32.3 per cent of of the media, to recognize the importance the programs reported having reached more of accidents and violent acts, in addition than 2,000 individuals during the 12 months to bolstering inter-sector actions pertinent leading up to the survey, whereas 64.9 per to the subject’ (Brasil 2001). Accordingly, cent stated that different types of interven- strategies are proposed such as Municípios tion were employed (MJ 2005). Saudáveis (Healthy Cities), Prefeito Amigo Overall, 34.5 per cent of programs had at da Criança (City Hall - Child Advocates) and least one partner institution, 32.7 per cent ‘Health Promoting’ Schools, the technical two or more partners, 19.6 per cent three dissemination of statistics on accidents and or more, and 13.1 per cent had no partner violent acts, support for seminars, vocational institution. Only 7.7 per cent of the programs and academic workshops, and incentives to focused on risk factors. Nonetheless, 28 per publishers of scientific journals to promote Art. 20, page 4 of 18 Beato and Silveira: Effectiveness and Evaluation of Crime Prevention Programs in Minas Gerais discussions and publish on these topics. The that provides victims and their families with Rule’s legal wording also expounds upon the the necessary medical, psychological and need to foster: social support.’ It also recommends specific care for female victims of violence, activities co-responsibility of citizens in a wider centered on the re-education of aggressors, sense, including their collective role support networks for care, such as shelters in upholding, providing input on and and units that provide legal abortions and monitoring this Policy. The strategy the physical and psychological rehabilitation also includes the promotion of safe of victims of violence. environments and healthy behav- Although the prevention activities envis- iors conducive to the prevention of aged are both fitting for the mission of the accidents and violent acts and their health area and technically pertinent, the respective risk factors, such as drugs preventive actions implemented remain in and alcohol abuse, firearms, non- practice limited, with the exception of the compliance with security rules, and pre-hospital assistance carried out by the interpersonal conflicts. These meas- rescue services and SAMU (Urgent Mobile ures encompass all domestic, leisure, Care Service), which works reasonably well work and transportation settings, as (Machado et al 2011). The programs for well as those involving the processes drug prevention and treatment are weak, of community living (Brasil 2001). leaving users who want to overcome their addictions with little or no support, or at The Ministry of Health further advocates that the mercy of clinics and farms, both private prevention actions be incorporated into all and affiliated with religious groups, which the assistance programs, plans and projects are not always trustworthy from a technical, for specific groups of the population and dif- ethical, and health standpoint. This situ- ferent health themes. ation has worrying implications for other The second directive deals with the moni- prevention programs, since addicted felons toring of accidents and violent acts, includ- have few alternatives for treatment under ing the training and mobilization of health the state public system and often become professionals working across all levels of SUS involved in violent and criminal activities care (Brazil’s Public Health System), given (Silva et al 2009). that improvement in the quality of informa- tion is deemed a priority. Prevention Programs in Minas Gerais The third directive concerns systematiza- One of the most comprehensive surveys of tion, expansion and consolidation of pre- the Minas case sought to assess the explana- hospital care so as to reduce the mortality tions proposed by scholars for the drop in of violent incidents, through promoting the the number of violent crimes in the state of organization and implementation of pre-hos- Minas Gerais. According to Betânia Peixoto pital care services, including the establish- (Peixoto et al 2011), the majority of the eval- ment of technical procedures, standardization uations carried out on the successes of the of equipment and vehicles for transporting Minas Gerais policy point to the institutional victims, and training of personnel. nature of the interventions performed. In The fourth directive addresses interdiscipli- particular, the development of mechanisms nary and intersectoral assistance to victims of of management was found to be key. This accidents and violent acts, through the struc- involved activities ranging from integration turing and organization of the network of among policies to the institutional partner- SUS services, to enable the system to better ships formed in projects such as Fica Vivo diagnose and treat users. The rule suggests (Stay Alive) and the IGESP (Integration and the deployment of an ‘interdisciplinary team Management of Public Security). This is yet Beato and Silveira: Effectiveness and Evaluation of Crime Prevention Art. 20, page 5 of 18 Programs in Minas Gerais another example of the Brazilian tendency for the incarceration and handling of juve- to focus on the implementation of public nile offenders to the management of fleets policies in a broader context, with no specific of police vehicles. Many programs that were focus on the different elements included in not directly conceived as part of violence these policies. prevention strategies have a direct impact, According to the review conducted, litera- since they involve actions that modernize ture hypothesizes that the factors contribut- the prison system, attenuating the level of ing to the success were: tension, which can lead to higher rates of recidivism in the system. In this respect, one 1) The creation of an institutional struc- of the most important actions taken was the ture by the Secretary for Social Defense implementation of the APACs (Associations and related structures. The institu- for the Protection and Assistance of Con- tional innovation employed by the victs), which promote a methodology that government of Minas Gerais, which has gradually been adopted at new detention sought to pool police prevention centers (Muhele 2013). actions and prisons together under Curiously, although a ubiquitous item the same organizational umbrella, was in all manifestos of union and corporate one of the factors considered (Sapori demands, there has been no consistent eval- 2007). Actions included: uation of the impact of investments in this a. Integrated Management of the plan on crime or violence. Without doubt, Social Defense system improved working conditions and infrastruc- b. Management of the prison system ture are important to professional work, but c. Social prevention of crimes the benefits of this in terms of effectiveness 2) The joint mobilization of other actors have yet to be shown. a. Interaction among different public actors Integrated Management of Social b. Interaction between public actors Defense and the community The second group of structured projects 3) The process of ‘metropolizing’ the concerns the Integrated Management of strategies adopted Social Defense, and entails actions such as the Implementation of Videomonitoring This whole process of structuring actions (‘Live Eye’), Integrated Building Units (State was organized along several axes. The Gov- Secretariat of Civil Defense – SEDS), Train- ernment of Minas Gerais, for its part, lists ing Academy of the Social Defense System, five actions for structuring programs in the Strengthening of Social Defense Analysis ambit of Social Defense: 1) actions related to and Intelligence, Implementation of the provision of Social Defense infrastructure; 2) Integrated Command and Control Center, Integrated Management of Social Defense; 3) Management of the Integration Policy of the Minas Mais Segura (Safer Minas); 4) Aliança Social Defense System, Rural Security, Inte- pela Vida (Pact for Life); and, more recently, grated Criminal Investigation (Civil Police), 5) actions associated with the World Cup. and Model Police Precinct (Civil Police). Sev- The most relevant axes to the present discus- eral more solid evaluations of actions are sion are addressed below. available, particularly for a previous version of the Integration of the Social Defense Sys- Infrastructure tem, which unequivocally showed the effec- The set of actions aimed at the provision of tiveness of this type of action (Soares 2011). an adequate infrastructure for the imple- Created in the middle of 2005, the manage- mentation of public policies ranges from ment model ‘Integration and Management the construction of buildings and facilities of Public Security’ (IGESP) constitutes the Art. 20, page 6 of 18 Beato and Silveira: Effectiveness and Evaluation of Crime Prevention Programs in Minas Gerais central plank of the policy for integrat- Other actions, such as those related to ing the military and civil police forces of videomonitoring, were assessed in a rather Minas Gerais. This comprises a management unscientific manner, being limited to the model based on the COMPSTAT (Computer temporal comparison of the rise and fall of Statistics) police management system imple- indicators, with no concern for the analysis mented in New York during the early 1990s of control areas or of confounding variables. and subsequently adapted to cater for local The actions for strengthening and analyz- needs by the Center for Studies on Crime ing intelligence were also never properly eval- and Public Security of the Federal University uated owing to implementation problems. of Minas Gerais (CRISP/UFMG 2008). The The first of these concerned the Integrated specificity of the IGESP lies in the systematic information system, the complexity of which interlocution and integrated action between required much more intensive training of the various different agencies of the social those responsible for data entry—training defense system. For instance, there is the which appears to have been provided at only qualified identification of recidivist felons a preliminary level. In addition, a major short- in specific integrated areas and of prob- coming in the ability to analyse the data was lems that go beyond the limits of these evident, which led to shortcomings in diag- areas and that permeate the city, such as car noses and in crime analysis reports. Some theft and robbery, assault and robbery of of these difficulties can be attributed to the individuals, and the involvement of adoles- obstacles imposed by State Secretariat for cents in numerous illicit activities. This con- Social Defense making it difficult for research- sists of an evidence-based, results-oriented ers to access data and its utilization (Castro policing experience. et al 2003; Jornal O Tempo 2014; Tribuna de Analysis of the results for 56 cities in which Minas 2012). the model was found revealed a correlation Furthermore, there is a growing number between the introduction of the IGESP and of reports from numerous press and research a reduction in the number of crimes perpe- bodies, highlighting the low quality of data trated. An evaluation of the impact of this and even claiming information tampering model was performed based on the use of by the agents overseeing the system (Castro regression models and of a control for the et al 2003; Jornal O Tempo 2014; Tribuna de implementation period in question (Ferreira Minas 2012). 2011). The results showed that, based on Recently, actions such as the implementa- conservative estimates, there was a 24 per tion of the Command and Control Center, cent reduction in property crimes and a 13 heavily funded by the Federal Government, per cent drop in personal crimes. Some evi- have been the focus of investigations at dence also points to a delayed effect of the the behest of various Ministries including model on homicide rates (Ferreira 2011). the Ministry of Sports and of Science and Recently, corporate rivalries have affected Technology, in a process that remains ongo- the results attained, a factor which has cer- ing and inconclusive. tainly contributed to the alarming rise in violent crime seen in the last three years. Safer Minas Gerais Following the subsequent deterioration of The third group of projects bears closer the program, no further outcomes evalu- resemblance to those typically conceived ations have been conducted. This case of as social prevention projects. The group warrants a more in-depth study of the evalu- includes the Fica Vivo (Stay Alive) program, ations of the processes in a bid to identify the Implementation of Centers for Crime the various challenges faced by the program, Prevention and the Police for Community. which ultimately led to a total distortion of Special attention should be given to the the original conception. Stay Alive program, in that it ranks among Beato and Silveira: Effectiveness and Evaluation of Crime Prevention Art. 20, page 7 of 18 Programs in Minas Gerais the most exhaustively assessed prevention schools, the police, community-based asso- programs in Latin America, having been the ciations and other forums within the com- subject of numerous monographs, theses and munity. The eligibility criteria defined by the dissertations, as well as reports and papers. State Secretariat for Social Defense to host This is partly due to its having its roots in the program are based on the existence of the academic world and subsequently being high rates of homicide in the community, transformed into public policy (Beato 2005; with the program´s target population com- Andrade & Peixoto 2006; Matta & Andrade prising young adults aged between 12 and 2006; Silveira 2007; Peixoto 2008; Silveira 24 years (Silveira 2007). 2008; Silveira 2010). It is important to point out that, despite being presented as a prevention program, Fica Vivo (Stay Alive) Stay Alive has never explicitly outlined Stay Alive is a program in the mold of the the risk factors that make up its mandate ‘Weed and Seed’ strategy that incorporates (Silveira 2007). Several risk factors can several elements from Boston’s ‘Ceasefire’ be deduced by interpreting the project program (Braga et al 2001; US Department it financed together with the Ministry of of Justice 2005; CCDP 2005). The program Justice (SEDS 2003) for expansion of the entailed the devising of supplementary Program in Belo Horizonte and the metro- Qualified Enforcement actions via a network politan region as of 2003, and by gleaning of institutions such as the civil and military from other documents produced by those police forces, the Ministry of Justice and running the program, namely: criminal judges, and via a Social Protection Network comprising educational activities, • Social vulnerability (as measured by in- professional training, and the use of schools dicators such as rate of entry to the for- and support for local business ventures. The mal job market, educational level, local program is aimed at gang-affiliated and other access to essential services, mortality at-risk youth. rate, teenage pregnancy, housing qual- Given the program’s aims and the way it is ity, etc.); implemented, Stay Alive can be considered a • Presence of a large number of youth broad (comprehensive) and focused project without proper supervision by adults at the same time (Silveira 2007). and/or involved in gangs; The program is broad since it fosters the • Low capacity of the community to exert implementation of projects and actions social control and mobilize external re- aimed at reducing the different risk factors sources for it. for crime existing in the community, and the strengthening of other protective factors The promise of the program lies not only in against crime. The program is community- the fact that it is inspired by community- based since the actions implemented are based programs, implemented at other not only aimed at, and implemented by, the sites (Braga et al 2001; US Department of community, but are essentially devised and Justice 2005, Community Capacity Devel- executed together with the community, who opment Office, n/d; Morrison, n/d), but thus play a key role in developing local plans also in its acknowledgement of the pivotal for public security, executing certain actions importance of the local community in the and monitoring the program. In addition, prevention of crime through informal social the program draws on different local actors, control, social cohesion and trust among seeking to form partnerships among them for neighbors (Banco Mundial 2003; English et the implementation of prevention actions. al 2002; Sampson et al 1997) and also in the Stay Alive is implemented by different actors importance placed on primary and second- in a range of settings, with an emphasis on ary prevention activities whose main target Art. 20, page 8 of 18 Beato and Silveira: Effectiveness and Evaluation of Crime Prevention Programs in Minas Gerais audience is the younger population (Mihalic employed the Difference-in-Differences et al 2001; Mihalic et al 2004). Underpinned method with Propensity Score Matching to by an alternative theoretical approach, analyse census sectors of shantytowns in which views crime as a common event per- which the Program was implemented. For vading everyday activities, the occurrence Difference-in-Differences, two samples not of which requires the convergence of a receiving the program were used: the city as motivated actor, a target, and the absence a whole and a similar group of shanty towns. of guardians of this target (Cohen & Felson The results show that between 2004 and 1979), it further reiterates the importance 2006 the program reduced 10.72 homicides of the community in prevention, since this per one hundred thousand inhabitants per is also centered around everyday activities semester less in the pilot area in relation to and whose emphasis is informal social con- the areas of comparison. The cost of a homi- trol (Souza 2013). cide avoided by the Stay Alive varies between A number of different types of evalua- about 93 and 112 thousand dollars. The rate tions of the program have been carried out. of return of the program is favorable for prac- One such evaluation, by Andrea Silveira tically all parameters utilized, varying from a (2007), sought to describe the different small return tax, 4 per cent, to a large return procedural components contained in the tax, 840 per cent. These results suggest that program. Silveira et al (2010) performed the Stay Alive Vivo present a favorable cost- a quasi-experimental study involving benefit ratio. analysis of time-series of the occurrence This type of networked action provided the of homicides in the shantytown of Morro inspiration for the integration programs of das Pedras in Belo Horizonte, Minas Gerais the IGESP. state, between 2002 and 2006. The number of homicides committed in the town was Projects involving other areas compared against those recorded in other The most important project underway violent and non-violent shantytowns and involving other areas is the Escola Viva, within other neighborhoods of the city, at Comunidade Ativa - EVCA (Living School, each of the Stay Alive Program´s phases. In Active Community), which has been rigor- order to test the hypothesis that the reduc- ously assessed. tion in homicides resulted from the actions implemented by the Program, a statistical Escola Viva, Comunidade Ativa model was built based on generalized linear – EVCA (Living School, Active models. In the first six months, a 69 per cent Community) reduction in the mean number of homicides This project was set up by the Secretary of was observed. In subsequent periods of State for Education of Minas Gerais with a fluctuations and partial resumption of the two-pronged objective: reducing violence Program, the effect of reduced homicides and improving school performance in vul- decreased, but the difference among coef- nerable areas. An experimental, non-rand- ficients compared with that of the initial omized evaluation was performed involv- period did not reach statistical significance. ing two groups: one group comprising 79 Even with full Program implementation, the schools participating in the project and a effect continued to be similar to the previ- second group of 21 schools serving as a con- ous periods, probably because the program trol (Corrêa 2007). was implemented in other violent slums in This project was run in the context of sec- the city. ond-generation administrative reforms and Another study on effectiveness conducted sought to go beyond the economic and fiscal by Betânia Peixoto (Peixoto et al 2008) reforms in an effort to introduce elements Beato and Silveira: Effectiveness and Evaluation of Crime Prevention Art. 20, page 9 of 18 Programs in Minas Gerais of accountability and foster development in all other variables (financial resources, school public state policies. In the field of educa- size and school performance). tion, the aim was to intervene in cities with This constitutes one of the few quasi- a lower Human Development Index (HDI) experimental evaluations carried out regard- and greater Social Vulnerability Index (SoVI). ing public security policies. In fact, Stay Alive, The project was formally created in 2003, via IGESP, and the EVCA project are among the Resolution SEE/MG nº 416/2003. The docu- few structured projects that have been sub- ment outlines the project goal as ‘better pre- mitted to rigorous evaluation of results. paring public schools to cater to the needs of children and young adults most impacted by Using the Maryland scale for the phenomena of violence and social exclu- projects and programs evaluations sion and providing the stability and condi- One approach for organizing the programs’ tions required for the learning process’ (SEE/ evaluations more systematically is through MG 2004). the use of the Maryland Scale (Sherman et al The Program’s target schools and con- 2006). Using this instrument, the performed trols were selected based on the Social evaluations are placed along a scale with levels Vulnerability Index – SVI1 and criminal occur- ranging from one to five described as follows: rences affecting the integrity of the school and the school community members (Westin Level 1 et al 2007; Soares and Andrade 2006). Correlation between a prevention program Comparison of the results for the two groups and a crime measure at some point in time. (see Table 1 and 2) revealed an increase in both, but the increase in the control group Example: Areas running the ‘Live Eye’ was found to be 20.9 per cent lower than in project are less violent than those the other schools, even after controlling for which are not.

Socioeconomic level Number of Schools in the No. of Schools in Control Group EVCA at 25% -0.11 to 0.03 16 4 0.04 to 0.18 37 8 0.19 to 0.33 17 6 Above 0.34 9 3 TOTAL 79 19

Table 1: Number of schools by socioeconomic level and group. Source: (Correa 2007).

1999/2000 2005/2006 School size Mean/100 N Mean/100 N Difference students students Project 5.5 73 6.3 75 14.50% Control 4.8 19 6.5 19 35.40% TOTAL 5.3 92 6.3 94 18.90%

Table 2: Results by school size. Art. 20, page 10 of 18 Beato and Silveira: Effectiveness and Evaluation of Crime Prevention Programs in Minas Gerais

There are problems regarding the inter- Many deem this to be the minimum accepta- nal validity of the affirmation, since this ble level for an evaluation. It considers many causal relationship cannot be established. of the problems of internal validity, includ- It is possible, for instance, that these areas ing history, maturation/trends, instrumenta- are populated by individuals who are more tion, test and mortality. The main problems predisposed to exercising some degree of include the effects of the selection employed control over events in their neighborhoods, in the model, as well as the regression to where this has a greater influence than the the mean, given the non-equivalence of the program itself. experimental and control conditions. One of the evaluations of the IGESP carried out Level 2 by the Catholic University of Measures of crime before and after the pro- (PUC-RJ) attains this level. gram, with no comparable condition to serve as control. In many cases, a control group Level 4 will be used, although researchers cannot Measures before and after the program in demonstrate comparability with the experi- multiple experimental and control units. mental group. Example: Crime measures decrease Example: Crime fell after implement- when comparing various comparable ing the Live Eye cameras. Compari- areas running the Live Eye with those sons between areas with the Live Eye in which the program was not tested. versus those in cities without it, show that crime declined. This design has a much better design of sta- tistical control of the extraneous influences There is a causal relationship, but the inter- on results. Only one of the evaluations out of nal validity cannot be affirmed. Although the various programs, performed for the Stay there may be a causal nexus between imple- Alive by the World Bank attained this level mentation of the program and the results, of analysis. other factors were not controlled for, such as the decline in crime in other areas with- Level 5 out the program, while other variables may Random choice of the areas to host the pro- have contributed to the decline (e.g. a more grams and of the control areas. visible police presence). Also, the terms of comparison were not explicitly defined. Example: Randomization of the areas Some authors hold that levels 1 and 2 should to receive the Live Eye and those to not be regarded as valid evidence (Cook and serve as controls. Campbell 1979). This level has the highest internal validity. Level 3 The following table (Table 3) includes Measures of crime before and after the pro- some, but not all, of these studies. Only those gram under experimental and compara- studies deemed relevant to the scope of this ble control conditions. The result is in fact paper are listed. derived from a regression model in which all variables involved are controlled. Conclusions One of the major differences between Bra- Example: Crime fell in an experimen- zil and the United States, in terms of gov- tal area running the Live Eye program, ernmental actions in the area of security, is but no decrease was observed in com- that the urgency of Brazil’s problems compel parable control areas. policy-makers to adopt multidimensional Beato and Silveira: Effectiveness and Evaluation of Crime Prevention Art. 20, page 11 of 18 Programs in Minas Gerais 2 1 4 1 1 Contd. Scale Maryland Method used Observation, interviews, descriptive analyses and regression model Observation, interviews, descriptive analyses Control group: Interviews, time-series Time-series - - - Type Process and out come Process and out come Outcome Process and out come Outcome - - - Objective Qualitative, compara tive case study Quasi-experimental ele ‘understand’ To ments of the preven tion program Assess Live Eye impact on BH city center - - - Title manage public order To and application of the law: forms of policing in a perspective comparing States – An Brazil-United analysis of the experi ences of Belo Horizonte – MG and Washington- DC Public Security in Brazil: challenges and perspec tives A contribution to crime prevention Crime prevention programs: from social processes to innovations Experi in Public Policy. ence of Stay Alive! Live Eye Project, the iris of the eyes public security: a geographic analysis. Program Fica Vivo (Stay Alive) The whole Social Defense system of MG. Fica Vivo (Stay Alive) Fica Vivo (Stay Alive) Live Eye Instit UFMG Ed. FGV UFMG FJP PUC MG Text PhD thesis PhD thesis PhD thesis Masters dissert Masters dissert 2007 2007 2008 2008 Year - Author PhD theses Silveira, Andrea Sapori, Luis Flavio Peixoto, Betânia Master theses Márcia Cris tina Alves Carvalho, E. A. Art. 20, page 12 of 18 Beato and Silveira: Effectiveness and Evaluation of Crime Prevention Programs in Minas Gerais -

4 2 Contd. n/a cable Scale not appli Maryland - Method used Control group Observation, interviews, descriptive analyses and Texts articles on the subject Interviews, questionnaires and observa tion Type Outcome Process Literature review Outcome - - - - Objective Evaluation with control group of the program Case study for implementation of program Assess explana tions of the rea sons behind drop in violent crime in MG, based on literature review Assess implemen tation of com munity policing in BH, identifying implementation difficulties - Title Evaluation of public policies for reducing school violence in Minas Gerais: The case of the Escola Viva, Comunidade Ativa (Living School, Active Community) Project. Fica Vivo (Stay Alive) case study Literature review (83 titles) on explanations for MG Reinventing the police: implementation of com munity policing - Program Escola Viva (Liv ing School) Fica Vivo (Stay Alive) BH Security Policy Community policing in BH BID Instit UFMG WB FJP Book Chap. Text Masters dissert Report Report 2011 2001 2007 2005 Year - Author Carvalho, Deborah Evaluation reports Beato Filho Peixoto, et al. and Papers chapters Beato Filho, Clau dio c. Beato and Silveira: Effectiveness and Evaluation of Crime Prevention Art. 20, page 13 of 18 Programs in Minas Gerais Scale Maryland 3 n/a 3 Method used Cost/benefit comparison of different types of program Interviews, official data, questionnaires Regression Model Type Outcome Process Outcome - Objective Evaluation of secondary primary, and tertiary pre vention programs Description of Citizen Security programs Modeling IGESP Impact - Title Evaluation of Cost/benefit of crime prevention and control programs in Brazil Citizen security in Brazil: Experience underway in Belo Horizonte. Citizen Security in the Americas- Active Investigation Project Organization and Informa tion in the Fight against Crime: An Evaluation of the Integration of Police in the State of Minas Forces Gerais, Brazil Program Fica Vivo (Stay Alive), Bolsa (Family Família Allowance), nas Uerê, Paz escolas (Peace in Schools), Proerd, Apac Community Advice, Fica Vivo (Stay Alive), Igesp IGESP Instit UFMG / WWC USA FJP Text Book Chap. Book Chap. Paper Year 2008 2007 2010 - Author Andrade, and M.V. Peixoto C/, Beato F., Silveira, A., Leite, K.R., Souza, E. Soares, Rodrigo and Vivei ros, Igor Table 3. Table Art. 20, page 14 of 18 Beato and Silveira: Effectiveness and Evaluation of Crime Prevention Programs in Minas Gerais strategies which act at different levels of affairs has led to a sharp increase in crime intervention. The preference in Brazil is for rates, to heightened fear and greater per- broader policies as opposed to programs and ceived risk among populations in the large projects, explaining the immense difficulties urban centers. The skepticism and disbelief in carrying out evaluations of outcomes as over the apparent impossibility of obtain- well as determining costs and benefits. The ing results is leading to a ‘banalization’ of outcome is that it becomes impossible to the phenomena of crime and violence, as if ascertain exactly what is working and what the public is doomed to live with fear and is not. insecurity. It is fair to suggest that, when it Latin America is distinguished from the US comes to crime prevention, our own igno- by the dearth of evaluations of the numer- rance on the subject ranks high (Beato 2012: ous crime prevention programs. Even areas Chap. IV). that have received attention, such as Bogotá, The specific impact of social policies and Cáli, Medellín, or Diadema in Brazil, have programs is even hazier, since the need for been subject to only superficial evaluation such projects is so pressing that any results of a highly preliminary nature (Beato 2002). attained, irrespective of the implications for The academic output and the documented the problem of crime, are associated with experiences centered on the crime preven- project success. No evaluations of these pro- tion programs and projects run in developed grams’ costs in terms of the results achieved English-speaking countries, particularly are available, nor on their effectiveness. What England and the United States, far outnum- is their effective impact on rates of violence ber those conducted in the developing coun- and crime? Which aspects work best? What tries of Latin America, where this output and timeframe is necessary for these to yield documenting of experiences remains scant. results? Which combinations are needed to Several characteristics contribute to this produce the most promising results? How imbalance. The first of these concerns the can unnecessary spending on well-meaning poor state of information systems in public but ineffective approaches be avoided? The security. All exploratory studies or literature analysis of these questions is of growing reviews on crime, violence, and the control importance, given the frequent shortage of policies in Latin America appear to either resources which governments at many dif- start or finish by highlighting the numerous ferent levels face, coupled with the natu- shortcomings in the information bases on ral tendency of identifying and reworking crime and violence. This is a serious situation crime prevention policies based on decisions which severely hampers performed studies derived from cost-benefit models. In sum- as well as policies, programs, and projects on mary, the challenge is how to generate data public security. The major challenge faced on the problems of security, how to trans- in crime studies across Latin America today form this data into information, and how to involves the information bases needed for translate this information into knowledge, advancing toward the attainment of empiri- which can provide a solid basis for actions cal proposals, as well as performing more that are ultimately assessable. sophisticated theory testing. Without this Another notable hindrance to progress is knowledge there can be no effective and the shortage of professional managers spe- meaningful actions. cialized in crime and other techniques with The implications of this situation for the the expertise needed to carry out diagnoses, design and evaluation of security policies implementation, monitoring and evalua- are clear. Policy-making in the crime and jus- tion of programs of this nature (Bengochea tice area is blindly based upon impressions, 2004; Silveira 2008; Soares et al 2009). Latin without guidance instruments. This state of America boasts very few specialists in crime Beato and Silveira: Effectiveness and Evaluation of Crime Prevention Art. 20, page 15 of 18 Programs in Minas Gerais or in the evaluation of public policies in the The specific case of Minas Gerais affords field. The academic clout of the vast major- another perspective on the growing divide ity of North American prevention programs, in recent years between universities and which invariably involve the collaboration research centers, which could well be con- of universities and research centers not ducting evaluations of the actions imple- only in the conception and execution of the mented. Despite being one of the states in programs, but also in their monitoring and which this partnership was most developed, assessment, is rare in the Latin American budget restrictions together with corporate milieu. Conversely, aversion and distrust still interests have contributed to a widening of prevail in these nations. this gap. However, assuming this remains a In the specific field of Brazilian criminol- strictly Brazilian problem is erroneous. In the ogy, few academic centers are dedicated to plenary presentation at the meeting of the the area, and even fewer study crime pre- American Society of Criminology in 1998, the vention and undertake empirical research Director of the National Institute of Justice into prevention. This also holds true for imagined a hypothetical and undesirable centers dedicated to public policies able to world in which the academic community generate applied knowledge focused on the was completely divorced from the interests evaluation and devising of programs and of the practical nature of the criminal justice projects. Although researchers dedicated to organizations: the study of violence in general do exist, the studies they produce are spread amongst a I suppose we could imagine a world number of different disciplines such as the in which the academic community fields of health, education, law, sociology, and the practice community did not political science, anthropology etc., with communicate, except to criticize no specific focus on prevention programs each other. We could imagine a world (Minayo 2003; Cano 2006). This results in in which the academic community the publication of low academic output goes about its research, maintaining that is removed from the practical imple- appropriate distance from the nitty- mentation of public policies. Again, this is gritty world of practice, publishing in stark contrast with the North American in academic journals, and criticizing academic scenario, which boasts a high vol- the crime policies of the country. In ume of published papers, books, bulletins, this same world, we could envision a research reports and sites on the internet community of practitioners who go describing experiences, providing manu- about their business, doing what they als for implementing prevention programs, thought best, criticizing the aloofness and disseminating results of evaluations of of academe, and ignoring the findings these programs, which evidences a strong of research. Yet we clearly do not want link between academic research and public to live in this world –we want to see policy making and management. some interaction between research Another important aspect of academic and practice (Travis 1998). underdevelopment in Latin America is the scant amount of public resources available Inadvertently, the director ended up paint- through government research funding bod- ing a vivid portrait of the current situation ies and agencies responsible for implement- in Brazil. Apart from some advances involv- ing public security policies, as well as private ing NGOs and a few research centers dedi- institutions dedicated to financing studies cated to this theme, the advances made are and commissioning studies on crime preven- extremely limited. At present there is no tion and evaluations. pipeline of studies, evaluations and experi- Art. 20, page 16 of 18 Beato and Silveira: Effectiveness and Evaluation of Crime Prevention Programs in Minas Gerais ments involving the police and these bod- forged, paving the way for expansion of this ies. The relationship between the academic type of strategy. world and the practical world of implement- ing public policies in the area of justice Author Information deserves its own chapter. Tensions between Claudio Beato is Professor of the Depart- these communities on issues such as the ment of Sociology at the Federal University economy, health or education. of Minas Gerais (UFMG) and coordinator of Another possible scenario is worth not- the Center for Criminality and Public Safety ing—one which Minas Gerais shares with Studies (Centro de Estudos de Criminalidade other States; namely, the growing ’corporati- e Segurança Pública – CRISP/UFMG). Andrea zation’ of public security as practiced by the Maria Silveira is Professor at the Faculty of police forces. There are several mechanisms Medicine of the Federal University of Minas by which this corporate isolation manifests Gerais (UFMG) and under-coordinator of and leads to a dearth of solid evaluations the Center for Criminality and Public Safety of effectiveness. The first and most impor- Studies (Centro de Estudos de Criminalidade tant is the corporate ownership of data and e Segurança Pública – CRISP/UFMG). information, justified by a certain ‘culture of secrecy,’ a remnant harking back to the Notes exception period during the military govern- 1 The indicator is related to population ments (1964–1985) (Bengochea 2004). The access to ‘five citizenship dimensions’: prevailing ideology of the intelligence com- Environmental (access to quality house munity that persists within the police forces, and sanitation); Cultural (access to and whose low degree of transparency schools); Economical (access to work and serves as a limited version of the ‘doctrine incomes); Juridical (access to juridical of national security,’ is now applied to pub- support); Security to Survival (access to lic security issues (Nóbrega 2010). There is a health services, alimentary security and growing consensus among scholars on the public welfare). subject that security is one of the sectors of the public administration that has least ben- References efited from the process of democratization of Andrade, M V and Peixoto, B T 2006 Cost the country. Over thirty years after the period Effectiveness of violence programs in Bra- of military governments, the sector remains zil. World Bank Report, n. 36525. Wash- wedded to dogmas which preclude the con- ington, D.C: World Bank. solidation of a greater degree of scientificity, Beato, C F 2012 Crime e Cidade. 1st ed. Belo transparency and accountability. Horizonte: Editora UFMG. Add to this the fear that, because the Beato Filho, C C 2002 Crime and Social Poli- implementation strategy was established at cies in Latin America: Problems and Solu- a broader and more systemic level of public tions. Woodrow Wilson Center Update policy, public administration as a whole will on the Americas. Available at http:// be held accountable for negative evaluations www.wilsoncenter.org/sites/default/ that might emerge. It should also be men- files/CitizenSecurity7.pdf. [Last accessed tioned that this is not a characteristic unique 18 Abril 2014]. to Minas Gerais, but is common to the hand- Beato Filho, C C 2005 Case Study “Fico Vivo” ful of States that have conceived and imple- Homicide Control Project in Belo Hori- mented public security policies. However, zonte. Washington, D.C.: World Bank. Minas Gerais is one of the states where this Beato Filho, C C 2008 Compreendendo e type of evaluation was systematically carried Avaliando Projetos e Programas de Segu- out and where pioneering partnerships were rança Pública: Ed. UFMG. Beato and Silveira: Effectiveness and Evaluation of Crime Prevention Art. 20, page 17 of 18 Programs in Minas Gerais

Bengochea, J L P et al 2004 A transição de Mihalic, S, Irwin, K, Elliot, D, Fagan, A uma polícia de controle para uma polícia and Hansen, D 2001 Blueprints for vio- cidadã, São Paulo Perspec, 18(1): 119–131. lence prevention. Juvenile Justice Bulle- DOI: http://dx.doi.org/10.1590/S0102- tin, July 2001. 88392004000100015 Mihalic, S, Irwin, K, Fagan, A, Ballard, Braga, A, Kennedy, D M, Piehl, A M and D and Elliot, D 2004 Successful pro- Waring, E J 2001 Reducing gun violence – gram implementation: lesson from The Boston Gun Project’s Operation Cease- blueprints. Juvenile Justice Bulletin, July fire. Washington D.C.: National Institute 2004. of Justice. Minayo, M C S, SOUZA, E R de (Eds.) 2003 Cano, I 2006 Políticas de segurança pública Violencia sob o olhar da saúde: a infra- no Brasil: tentativas de modernização política da contemporaneidade brasileira. e democratização versus a guerra con- 1st ed. Rio de Janeiro: Editora Fiocruz,. tra o crime. Sur, Rev. int. direitos human, Morrison, S The city of Albuquerque, New 3(5): 136–155. DOI: http://dx.doi.org/ Mexico – weed and seed, best practices: 10.1590/S1806-64452006000200007 Evaluation-based series. Washington, D. Centro de Estudos de Criminalidade e Seg- C.: US Department of Justice, Community urança Pública 2008 O Que é o IGESP. Capacity Development Office. Informativo Crisp/UFMG: Crisp/UFMG. Muhle, E P 2013 A prisão terrena no paraíso Cohen, L E and Felson, M 1979 Social celestial: APAC, uma alternativa humana change and crime rate trends: a routine ao cumprimento da pena privativa de activity approach. American Sociological liberdade. Unpublished thesis (MSc) Review, 44: 588–608. DOI: http://dx.doi. Pontíficia Universidade Católica do Rio org/10.2307/2094589 Grande do Sul. Community Capacity Development Nóbrega Jr., J M P 2010 A semidemocra- Office- US Department of Justice 2005 cia brasileira: autoritarismo ou democ- Weed and Seed Implementation Manual. racia? Sociologias, (23): 74–141. DOI: Washington D.C.: US Department of http://dx.doi.org/10.1590/S1517- Justice. 45222010000100005 Corrêa, D 2007 Escola Viva e Comunidade Peixoto, B 2008 Uma contribuição para a Ativa. Unpublished dissertation (MA), prevenção da violência. Unpublished the- Faculdade de Educação (UFMG). sis (PhD). Cedeplar/UFMG. Ferreira, S G 2011 Segurança Pública nas Peixoto, B, Andrade, M V and Azevedo, J Grandes Cidades In: Bacha E, Schwartz- P 2008 Prevention and Control of Homi- man S Brasil: A Nova Agenda Social. Rio cides: An Evaluation of Impact in Brazil. de Janeiro: LTC. pp. 287–313. In: The Annual Meeting LACEA-LAMES Kelly, K D, Caputo, T and Jamieson, W 2008 on Rio de Janeiro. 2005 Reconsidering sustainability: Peixoto, B G, L S, Cruz, S F V and Sousa, some implications for community-based R 2011 Reflexões sobre a atuação gov- crime prevention. Critical Social Policy, ernamental na promoção da segurança 25(3): 306–324. DOI: http://dx.doi. pública (Belo Horizonte, 2003, 2010). org/10.1177/0261018305054073 Texto para Discussão 2/2011. FJP. MG. Machado, C V, Salvador, F G F and O’Dwyer, Rigakos, G 1996 New right, new left, new G 2011 Serviço de Atendimento Móvel de challenges: understanding and respond- Urgência: análise da política brasileira. ing to neoconservatism in contemporary Rev. Saúde Pública, 45(3): 519–528. criminology. Critical Criminology, 7(2): DOI: http://dx.doi.org/10.1590/S0034- 75–91. DOI: http://dx.doi.org/10.1007/ 89102011005000022 BF02461115 Art. 20, page 18 of 18 Beato and Silveira: Effectiveness and Evaluation of Crime Prevention Programs in Minas Gerais

Rolim, M 2006 Prevenção e Fatores de Risco 44(3): 496–502. DOI: http://dx.doi.org/ In: Rolim M. A síndrome da rainha ver- 10.1590/S0034-8910201000030001 melha, policiamento e segurança pública Soares, L E 2006 Segurança pública: pre- no século XXI. Rio de Janeiro: Jorge Zahar sente e futuro. Estud. av. 56: 91–106. Editores. DOI: http://dx.doi.org/10.1590/S0103- Sampson, R J, Raudenbush, S and Earls, 40142006000100008 F 1997 Neighborhoods and violent Soares, L E, Rolim, M and Ramos, S 2009 O crime: A multilevel study of collective que pensam os profissionais da segurança efficacy. Science, 277: 918–924. DOI: pública, no Brasil. Brasília: Ministério da http://dx.doi.org/10.1126/science. Justiça. Available at http://www.rolim. 277.5328.918 com.br/2002/_pdfs/300809.pdf [Last Silva, J et al 2009 Familiares e Pessoas Con- accessed 18 April 2014]. hecidas de Usuários de Drogas Ilícitas: Soares, R and Viveiro, I 2010 Organization Recorte de Opiniões Sobre Leis e Políti- and information in the fight against crime: cas Públicas de uma Comunidade da An evaluation of the integration of police Zona Oeste do Rio de Janeiro, Brasil. Rev. forces in the state of Minas Gerais. FJP. MG. Latino-Am. Enfermagem, 17: 803–809. Soares, R and Viveiros, I 2013 The pro- DOI: http://dx.doi.org/10.1590/S0104- gram of Integration and Management in 11692009000700008 Public Safety in Minas Gerais: Brazil. A Silveira, A M 2007 Avaliação do programa descriptive analyses. In: Cook P J, Machin Fica Vivo no morro das pedras em belo S, Marie O, Mastrobuoni G Lesson of Eco- horizonte. Unpublished thesis (PhD), Uni- nomics of Crime: What reduces offending? versidade Federal de Minas Gerais. Massachusetts Institute of Technology. Silveira, A M 2008 A Prevenção de Hom- Travis, J 1998 Plenary Presentation to the icídios: a experiência do Programa Fica American Society of Criminology. Avail- Vivo no Morro das Pedras, Revista Edu- able at http://www.nij.gov/about/ cação & Realidade, 33(2): 163–176. speeches/past-directors/Pages/asc.aspx DOI: http://dx.doi.org/10.1590/S0034- [Last accessed 18 April 2014]. 89102010000300013 US Department of Justice 2005 Weed and Silveira, A, Assunção, R, Silva, B and seed implementation manual. Community Beato, C 2010 Impacto do Programa Capacity Development Office. Washing- Fica Vivo. Revista de Saúde Pública, ton, D.C.: Department of Justice.

How to cite this article: Beato, C and Silveira, A M 2014 Effectiveness and Evaluation of Crime Prevention Programs in Minas Gerais. Stability: International Journal of Security & Development, 3(1): 20, pp. 1-18, DOI: http://dx.doi.org/10.5334/sta.dr

Published: 16 May 2014

Copyright: © 2014 The Author(s). This is an open-access article distributed under the terms of the Creative Commons Attribution 3.0 Unported License (CC-BY 3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. See http://creativecommons.org/licenses/by/3.0/.

Stability: International Journal of Security & Development is a peer-reviewed open access journal published by Ubiquity Press OPEN ACCESS