The quality of political leadership is a central determinant of state capacity. This column studies the prevalence and impact of candidates with criminal backgrounds holding public office in Brazil. Individuals who have been charged with a crime are twice as likely to run for office and to be elected, a pattern that persists across political parties and levels of government. Electing mayors with criminal behaviour in their past increases the incidence of political patronage and worsens public health outcomes such as underweight births and infant mortality.
The quality of political leadership is a central determinant of institutional performance and state capacity. Political parties select competent politicians depending on their incentives (Nannicini and Galasso 2009), with relevant consequences for public policy (Baránek and Titl 2024). On the other hand, institutional designs might matter for politicians’ behaviours, as rent-seeking is affected by monitoring and electoral behaviours (Daniele and Giommoni 2021).
A growing literature has highlighted the role of politicians’ personal characteristics – particularly integrity – in shaping public trust, bureaucratic quality, and the provision of public goods (Besley 2005, Dal Bó and Finan 2018, Daniele et al. 2023). In response, several countries have implemented legal restrictions disqualifying individuals with criminal convictions – or, in some cases, pending charges – from holding elected office (Transparency International 2016).
In a recent study, we provide first evidence on whether individuals with criminal charges are overrepresented among elected officials, focusing on Brazil. We then assess the causal effects of their election on policy outcomes. We combine individual-level administrative data on criminal charges, electoral participation, and public service provision to document both selection patterns and downstream consequences. Our findings indicate that criminally charged candidates are not only disproportionately represented in local politics but also have measurable adverse effects on public health outcomes when elected to office.
Overrepresentation of criminally charged individuals in politics
Using nationwide electoral and judicial records, we show that individuals with a history of criminal charges are significantly more likely to contest – and win – local elections than would be expected based on their population share. To ensure that the charges predate political careers and are not a consequence of holding office, we restrict our analysis to first-time candidates, thereby isolating selection into politics from potential legal exposure after election. Specifically, the probability of running for office is approximately twice as high for individuals with criminal charges relative to the general adult population (Figure 1). These patterns persist across multiple election cycles and are particularly pronounced at the municipal level, where monitoring capacity is often limited.
Importantly, many of these candidates face charges for crimes directly related to the abuse of public office, such as fraud, embezzlement, and corruption. This suggests that their participation in politics is not random but may be strategic, reflecting an attempt to access rents or avoid prosecution. Moreover, criminally charged candidates differ systematically from their non-charged counterparts along several observable dimensions. They tend to be older, predominantly male, and less educated.
Figure 1 The share of individuals with prior criminal charges in the entire Brazilian population (yellow bar) and among candidates and elected individuals (blue and red bars)
Policy consequences: Public health deterioration
To assess the impact of criminally charged politicians on policy outcomes, we exploit close mayoral elections between candidates with and without criminal charges. This empirical strategy allows us to identify causal effects by comparing municipalities that are ex ante similar but differ in the type of candidate elected. We find that municipalities governed by criminally charged mayors experience a significant deterioration in public health indicators. Specifically, these municipalities exhibit an 8% increase in the incidence of underweight births and a 21% increase in infant mortality rates.
These effects are not observed in such policy areas as education, where municipal governments have similar administrative responsibilities. This heterogeneity suggests that the health sector may be particularly vulnerable to political appointments and managerial decisions, possibly due to its organisational complexity.
Mechanism: Patronage in health sector appointments
Interestingly, we identify political patronage as a central mechanism driving the observed effects. Mayors with criminal charges are more likely to engage in discretionary hiring and appoint political affiliates to key administrative positions in the public health system. This behaviour is consistent with the use of public employment as a reward for political loyalty rather than a means of improving service delivery.
By disaggregating public employment data, we find that the increase in politically motivated hiring is most pronounced among high-ranking positions (e.g. hospital directors), where decisions have wider operational implications. This pattern supports the interpretation that even a small number of poorly selected appointments can have cascading effects on public service quality.
Electoral audits and strategic candidacy
We further investigate whether institutional oversight affects the likelihood that criminally charged individuals run for office. To this end, we leverage the randomised audit programme conducted by the Brazilian federal government, which publicly discloses findings of financial misconduct among municipal politicians (Ferraz and Finan 2008).
Our analysis indicates that audits conducted during election years lead to a significant reduction in the candidacy of individuals with criminal charges. This effect is not observed for audits conducted in non-election years, suggesting that the timing of exposure interacts with strategic candidacy decisions.
Implications for political selection and institutional design
Our findings contribute to the literature on political selection and governance by providing causal evidence on the consequences of electing candidates with criminal backgrounds. They also underscore the limitations of relying solely on legal disqualifications or episodic oversight to ensure candidate quality. While pre-election audits can deter some forms of entry, their effects are transitory and may be circumvented through strategic timing.
More broadly, the results suggest that weak political screening mechanisms – combined with low levels of institutional accountability – can generate substantial social costs, particularly in policy areas where mismanagement has immediate and severe consequences. Addressing these issues requires sustained investments in institutional capacity, transparency, and professionalisation of the public sector.
Source : VOXeu
