Public sector governance statistics in Brazil: Challenges and perspectives, and focus on transparency

1.Introduction

As it has been discussed for some time and more effectively since the creation of the Praia Group and the SDG (Post-2015 agenda), governance statistics11 are still often on the margins of regular production by official statistical offices. However significant progress is being made around the world.22 At the same time, this is still an area of exploration. Brazil is no exception.

In addition to classical definitions related to the proper functioning of central, federal, and local governments, such as the definition adopted by the OECD (“Governance is the exercise of political, economic and administrative authority necessary to manage a nation’s affairs”), governance is understood here as an umbrella concept encompassing many dimensions, including those listed in the recently published handbook [1].

Within the official Brazilian statistics office (Brazilian Institute of Geography and Statistics – IBGE), federative issues have been a trigger for data production efforts since the 1990s. In contrast, the idea of structuring governance as a field of study is quite recent.33

Even if statistics related to governance have been produced by some institutions since the last century, the discussion on the subject gained momentum in Brazil in the early 2000s, due to efforts to rationalize the functioning of the public sector, initiatives (or reaction from the public administration at the three levels of government – federal, state, and municipal) and actions made by control bodies. New laws are also noteworthy, such as the Fiscal Responsibility Law – 05.04.2000 – and the Access to Information Law – 11.18.2011 (this law was first proposed in 2003). Brazil had a transparency portal online in 2004 and was part of the Open Government Partnership (OPG) and its implementation in 2011 [3, 4]. At the beginning of the 2010s, efforts accelerated but seem to have lost momentum due to lack of focus and funding.

This contribution aims to present some aspects of the production of statistics related to governance in Brazil. In addition to this brief introduction, the article includes three sections. The first section outlines a brief history of the production of statistics related to the topic in Brazil. The second section focuses on IBGE statistical production. It provides more details about two of the main Brazilian governance data sources: the Survey of Basic Municipal Information (aka MUNIC) and the Survey of Basic State Information (aka ESTADIC). The third section includes an analysis using different investigations (data sources) on the topic of transparency, in order to articulate opportunities and challenges in this field of investigation. At the end, final considerations address perspectives for the topic in Brazil.

2.A brief history of governance statistics in Brazil

In Brazil and in the world, audit institutions stand out with a relevant role in defining criteria for (good) public sector governance. Likewise, they evaluate the compliance with laws in favor of transparency and other dimensions of governance. In this context, the Federal Court of Accounts – TCU, the Federal Public Prosecutor’s Office – MPF and the Ministry of Transparency and Controller-General of the Union – CGU stand out. In addition to these institutions, it is worth mentioning the Regional Center for Studies on the Development of the Information Society – CETIC.br, with research related to electronic government. Next, we will summarize relevant data production from each of these institutions.

2.1TCU (Tribunal das Contas da União; Federal Court of Accounts)

As part of its mission to improve public administration for the benefit of society through external control, in 2013, TCU published a guide related to Basic Governance principles. This guide included good governance practices that could improve the performance of public bodies and entities, based on three mechanisms: leadership, strategy, and control [5].

In a progressive effort, the Court has been studying several governance dimensions, culminating in an integrated system. Between 2013 and 2018, TCU conducted surveys on governance and management in four distinct themes. The Public Organizational Governance Survey has assessed the degree of maturity (good practices) in the governance and management practice of federal public organizations. The most recent questionnaire (sent to federal public bodies) included the following dimensions: 1. Leadership; 2. Strategy; 3. Accountability; 4. People Management; 5. Information Technology Management; 6. Contract Management; 7. Core Activity Results.

Another tool on the TCU website is the Fraud and Corruption Controls Weakness Index Map (https:// meapffc.apps.tcu.gov.br/). According to available metadata, 287 institutions of the federal executive power were evaluated to define how much each institution was exposed to the risk of fraud and corruption. To measure exposure to such risks, a set of controls was divided into five categories: appointment of managers; risk management and internal controls; ethics management and integrity program; internal audit; transparency and accountability. These five categories were then mapped and related to two characteristics that enable the calculation of the exposure map: (a) economic power, based on the General Budget of the Union and the Global Expenditure Plan; and (b) regulatory power, defined as the power to create rules or oversee the private sector. Figure 1 locates a set of federal governments in these bidimensional summaries.

Figure 1.

Mapping of Weakness of controls vs economic power and regulatory power. Source: https://meapffc.apps.tcu.gov.br. Note: Horizontal Axis: Fraud and Corruption Controls Weakness Index (0 – low …1 – very high). Vertical Axis (panel a): Economic power (10 million to 1 trillion, log scale). Vertical Axis (panel B): Regulatory Power (0–100).

This tool does not say to which year the data refer, but from the results it seems that the mapping was done from 2016 to 2018. For example, (a) highlights Petrobras (oil company – mixed economy company under state control) with strong economic power and low values in the Fraud and Corruption Controls Weakness Index. After a series of scandals involving this company in 2015/2016, one would then question the efficiency of an evaluation based on the presence/absence of structures/procedures. That is however a general deficiency of many such indicators that have no easy task to evaluate if governance mechanisms are working in practice inside a (public or private) company/administration.

3.IBGE main data sources on governance

As said before, IBGE approaches governance (as a broad concept) with some statistical operations. The main one is the Survey of Basic Municipal Information – MUNIC, started in 1999 and applied annually, except for Demographic Census years.66 Since 2012, the Survey of Basic State Information – ESTADIC has been reproducing the same investigation as MUNIC at the level of the states and the Federal District. Although there was no specific section or supplement with the name of governance before 2019, several themes concern governance statistics at the municipal and state level, such as:

All these themes concern dimensions identified with the effectiveness of administrations, the existence of an inclusive and transparent public service, with space for social/citizen control mechanisms, social participation, etc. In other words, they concern the governance of the public sector (in this case, municipal, state and district governments) and the evaluation of mechanisms for “good governance”.

Since the 2010s, IBGE has tried to organize MUNIC and ESTADIC to rotate themes every four years, and has been successful, despite budgetary constraints. The budget of IBGE (a public foundation) is influenced by decisions of the ministry to which it is subordinate, currently the Ministry of Economy, and many investigations depend on cooperation agreements with other institutions that will then have a say on the investigation content and transfer funds to IBGE. Table 4 in appendix shows the distribution of topics (themes) investigated in MUNIC (since 1999) and in ESTADIC (since 2012).

Regarding the methodology of these surveys, it is worth discussing the collection method. Until 2019, the surveys were carried out with a paper questionnaire, either applied by IBGE interviewers going to city halls and state executives or left for self-completion. This method varied according to the practicality perceived by each IBGE state division responsible for the data collection. The advantage of the paper questionnaire comes from the fact that the survey covers several themes in the same year, so the questionnaire could “travel” among the different administrative divisions. Challenges came from transport, typing, validation.

Since 2020, and accelerated by the pandemic, the collection method was changed to online questionnaires (Computer Aided Web Interview, CAWI). This method has always raised fears, due to the issue of non-response. Some small municipalities do not even have a landline.88Table 1 shows that, in 2021, 3,770 or 68% of the municipalities had up to 20,000 inhabitants, concentrating 15% of the population. In most cases, these municipalities have few resources and administrative structure. Regional inequalities in Brazil are significant.

IBGE researchers responsible for the implementation and analysis of MUNIC and ESTADIC have the perception that these surveys are poorly under-used considering their potential (for monitoring government plans, policies, assessment of institutionalization levels of policies/themes, benchmark studies with comparison between federative entities, etc.). This situation is more a pity since the results are available in open data: the databases are available for free in xls/ods format on the IBGE website among other dissemination initiatives by IBGE.

These Brazilian (establishment) surveys coincide with instruments existing in other Latin American countries. In Mexico, government censuses99 started in 2009, with different contents, initially investigating justice and security sectors and then other governments offices. This investigation includes a set of statistics with the purpose of supporting the processes of design, implementation, monitoring and evaluation of government public policies. In Peru, the National Registry of Municipalities, established in law since 2001 [11, 12], also investigates issues related to different policies in all the municipalities of the country.

Besides ‘government censuses’, Mexico (INEGI) draws up a benchmark of the production of public sectors governance statistics. The research agenda includes ‘social cohesion and delinquency prevention’ [13], ‘quality and government impact’ [14], ‘Regulatory quality and government Impact on companies’ [15], ‘victimization of companies (including corruption)’ [16]. INEGI has also had an important role in the production of a manual about corruption surveys, coordinated by UNODC and UNDP.

Peru (INEI) also has a consistent production, regarding ‘governance, citizenship perceptions, democracy and trust in institution’ [17], integrated inside a household survey about poverty and living conditions [18]; see also Herrera et al. in this issue.

Concerning other IBGE data sources, it is worth exploring its household surveys. The National Household Sample Survey – PNAD was, between 1967 and 2015, the main source of socioeconomic data for Brazil. It was a face-to-face household cross-sectional multipurpose survey.1010 PNAD focused on the labor market but included as well specific modules (which varied year after year, since they were commissioned by different ministries, after technical agreements and resource transfers).

Since 2012, a new survey (Continuous PNAD) has started with a different logic. It is a survey under a rotating panel format, in which households respond 5 times, once per quarter, before leaving the sample.1111 This investigation replaced the PNAD (annual, national) and the Monthly Employment Survey (PME, 1980–2016) which covered only the 6 main metropolitan areas). Continuous PNAD provides national estimates with more geographic disaggregation than the previous PNAD and PME. Employment statistics are disclosed by moving quarters. The research also includes modules, investigated throughout the year, such as for housing, but also in specific quarters, such as education (2nd quarter) and ICT (4th quarter).

Inequalities of access to public services have been part of investigations in PNAD, Continuous PNAD and other surveys such the National Health Survey (PNS – 2013, 2019), fielded by IBGE in partnership with (and financed by) the Health Ministry. PNS is a necessary next step further (health was a module in PNAD in 1998, 2003 and 2008). The need for detailed information led to the implementation of a single purpose survey. PNS interviews can be longer than 1 hour.

Issues related to democracy, political participation, trust in institutions and civic engagement (themes closer to the core definition of governance) were investigated with PNAD in 1988 (close to the end of the dictatorial military regime in 1985) and never since.

Victimization has always been a sensitive and very relevant topic in Latin America, the area of the world with the highest homicide rates. This theme was investigated in PNAD modules in 1988 and 2009. For years, high turnover in ministerial departments and changing political priorities are among the reasons for difficult negotiations between IBGE and possible sponsors (Ministry of Justice, Justice National Council, etc.). In contrast, a new victimization (and access to justice) module was commissioned by the Ministry of Justice and has been fielded with Continuous PNAD in 2021 and 2022.

Modules regarding Information Communication Technology (ICT) in PNAD (2003, 2008, 2013) and in the Continuous PNAD (annually since 2016), commissioned by the Ministry of Communications, have been linked mainly to infrastructure issues (existence of ICT equipment at the households, access to cabling, etc.). The population’s access to e-government services was investigated only in 2003 and 2008. As stated, investigations concerning the main dimensions of governance, as defined in the Praia Group Handbook [1], with persons or households as collection units are somewhat rare in Brazil. More details about perspectives will be mentioned in the closing remarks. Table 2 summarizes the main IBGE governance data sources at the household level.

4.Approaches of transparency at the municipal level

After having explored governance data sources from different producers (Section 1) and from IBGE (Section 2), here is an example of a study that uses transparency data at municipal level.

Figure 2.

Mean value of indexes based on surveys about transparency, by region and population size of municipalities (2014–2016). Source: [17, p. 206]. Notes: 1. Because of the different number of questions used to build each index, they were normalized to 100 to allow for comparison; 2. Color scale (red, orange, yellow, green) is defined in each column, from lowest to highest transparency. 3. Based on Survey of Basic Municipal Information – MUNIC (IBGE, 2015); ICT Electronic Government survey 2015 (CGI.br, 2016); National Transparency Ranking (Ranking Nacional de Transparência – RNT), 2nd edition (MPF, 2016); and Brazil Transparency Scale (Escala Brasil Transparente – EBT), 3rd edition (CGU, 2016).

Transparency (part of the openness dimension) is one of the main dimensions of governance and probably that with most data available in Brazil. This section retrieves then part of a previous analysis [19], which illustrates the possibility of combining different data sources. Here we try to pinpoint regularities and again additional insights from the data (triangulation logic).1212 The exercise was based on four indexes, created by the author using surveys questions from four data sources:

All four indexes were a sum of positive responses to the existence of the above-mentioned transparency mechanisms. As for results, Fig. 2 shows a territorial distribution of transparency levels in the four investigations (unit of analysis is the municipal executive branch). The North and Northeast regions of Brazil are the ones with the lowest standard of living, showing lower transparency levels in rankings. The most relevant segmentation regarding levels of transparency of the municipalities is however constructed with population size of the municipalities. The larger municipalities are more structured with instruments related to transparency.

Beyond differences of questions, other challenges to the comparison can be highlighted. First, there were methodological differences (some already pointed out in Sections 1 and 2), such as the fact that some sources are the census and others are based on a sample of municipalities. Secondly, CETIC.br considers that municipalities are respondents under imperatives of confidentiality, sharing only aggregate data, not even with a cooperation agreement and guarantee of confidentiality with IBGE, which was negotiated at the time of the disclosure of the ICT Gov 2017 survey results. Data matching at the individual level (municipality by municipality) was then impossible.

In addition to the methodological differences and diverse understandings of confidentiality, the exercise shown above presented an additional difficulty. CGU data did not include municipality codes produced by the official Brazilian statistical office (IBGE). To better explore the data (merge databases) it was then necessary to edit the names of the municipalities and cope with related difficulties (names with spaces, accents, apostrophes, municipalities that have changed their names over time etc.).

5.Final considerations

In this article we present some data sources and efforts related to the production of governance statistics in Brazil, both by the official statistics agency, IBGE, and other public administrations, all linked to investigation and control purposes or the mapping of ICT in Brazil.

IBGE has a consistent production, especially on issues related to the capacity (resources, policies, etc.) of states and municipalities to fulfill their missions (MUNIC and ESTADIC surveys). Household surveys have been investigating governance issues only marginally and irregularly, mostly about inequalities of access to public services. IBGE company surveys regularly collect a series of information, but in many cases the respondent is an external accountant. That means a challenge to the investigation of issues related to practices (management, for example) for which only the top management or the company’s human resources department could properly report.

The theme of victimization of companies due to corruption (important for the monitoring of target 16.5 of the SDGs) then collides with the current structure of investigations. To advance on this research agenda, it will be necessary to innovate in terms of the structure of the surveys, with investment and tests. The production of indicators related to the connection between citizens and public services (and their quality) is also part of IBGE’s production, but it depends on agreements with Ministries (therefore political decisions) to finance the statistical operations.

In relation to other producers mentioned (CGU, MPF and TCU), there has been no recent update (since 2016 or 2018) of the studies. Specific study would be necessary, but the situation may be linked to the fact that Brazil went through its worst economic crisis in 2015 and 2016, with a slow and insufficient recovery since then. The crisis coincided with the impeachment of President Dilma Roussef (August 2016), followed by governments with high top management turnover and fiscal austerity.

Collaboration and standardization in the Brazilian statistical system could be fostered with the update of the main statistical laws (from the 1970s). Problems of data sharing are one of the major challenges, which lead to inefficiency and reengineering. One can also mention the need to invest more in promoting meetings, seminars, training of data producers to adapt them to statistical purposes, to facilitate the merging of database, etc.

The exercise of bringing together four data sources that address transparency at the municipal level disclosed some of the difficulties faced by researchers due to the lack of standardization. At the same time, results showed consistency in the approach, as the most populous municipalities have a more developed administrative structure and a structure of websites including services for e-government transactions and transparency portals (some evolving as open data projects) along with other transparency mechanisms.

In Brazil there is clearly a “lack of governance” regarding governance statistics, i.e., there is no institutional structure to orchestrate the data/indicator production. The official statistical body has a proactive role, but is faced with budgetary and personnel limitations, which make it difficult to build a comprehensive and regular research agenda.

Inter-institutional initiatives have historically brought results: surveys, studies, training of civil servants, compilation of statistics and publications, but they suffer from staff turnover and, with exceptions (e.g., ENCCLA – National Anti-Corruption and Money Laundering Strategy), do not usually reach a “critical mass” necessary to institutionalize themselves. At the same time, the participation of some organizations in international forums and the requests made in these forums and expressed by the international system also lead to data collection and/or the production of studies, however without guarantee of continuity.

In terms of new research fronts, the agenda related to companies’ victimization (including by corruption) may be promising avenues due to the interest of employers’ associations. External financing has been one important driver to innovation in the Brazilian Statistical System. Regarding perceptions (especially the trust dimension), there is still a long way to go, which may be accelerated if Brazil moves on in the project of integrating the OECD. In 2022, this country had adhered to about 44% of instruments.