Reforms to selection in public organisations are emphasised by the World Bank as an important aspect of its development programs (Evans 2008). While civil service systems have been introduced in most countries, temporary contracts, job categories exempted from formal selection rules, and political influence on the outcomes of public examination are still ubiquitous across bureaucracies, leaving ample pockets of discretion to political powers (Evans and Rauch 1999, Grindle 2012).
Civil service systems characterised by objective and meritocratic personnel policies are meant to keep public sector jobs from being used as patronage tools. But if politicians retain discretion in public employment decisions, jobs can be allocated to reward political supporters rather than to the most competent candidates. A counter-argument often put forward in this debate is that some discretion may be needed to ensure that the best overall candidates are selected. Discretion, this argument goes, allows politicians to hire and promote candidates whose competence and motivation are not easily measured in formal examinations.
The extent to which discretion is used as a patronage tool in modern bureaucracies is the missing piece in this debate. In a new paper (Colonnelli et al. 2019), we ask how often discretion in public employment decisions is used as a patronage tool, and the consequences on the selection process in the context of the Brazilian public sector. Most of that sector operates under a civil service system, but Brazilian politicians retain a significant amount of de facto discretion in public employment decisions (Grindle 2012). How do they use this discretion?
To answer this question, our analysis exploits a unique dataset of roughly 2,000,000 supporters of Brazilian local parties, namely local politicians or campaign donors. These individuals can be clearly identified as supporters of a specific local party based on their party affiliation or the recipient of their donations. We match these supporters to administrative data on the universe of private and public sector workers over the period from 1997-2014. This allows us to track the labour market careers of supporters of Brazilian political parties. Since our data covers all Brazilian public sector workers, we can analyse the role of political connections at all layers of the public hierarchy, from high level bureaucratic positions, to the bureaucracy’s middle-tiers, to jobs as frontline providers.
In the first step of our analysis, we ask whether individuals who support the party in power in a municipality enjoy easier access to public jobs. To isolate a causal link between an individual’s political connections and her public sector career, we compare supporters of the winning party in a municipality to supporters of the losing party in the same municipal election. We focus on municipalities where the winning party (the party of the elected mayor) had a very small margin of victory over the runner-up. As the outcome of these elections was close, it is unlikely that supporters of the two sides differed significantly before the election – an intuition that we formally confirm in the data using observable pre-election characteristics of the supporters. In the years leading up to the election, these supporters have a similar probability of being employed in the public sector. After the election, individuals with connections to the party in power are significantly more likely to have a public sector job relative to those connected to the losing party. These patterns are clearly depicted in Figure 1, which shows the average public employment numbers for supporters of the mayor and for those of the runner-up in the years around an election. Supporters of the two sides have the same public employment probability before the election, which takes place at time 0, indicated by the vertical dashed line. Immediately after the election, the public employment probability for supporters of the elected mayor spikes, while probability for supporters of the runner-up declines. Among local politicians (top panel), we see a post-election gap of 12.4 percentage points; among campaign donors (top panel), the gap is 6.7 percentage points.
Several patterns suggest that patronage is a relevant explanation behind these patterns of preferential hiring. First, we find that political connections play an important role across the public sector hierarchy, not only for high level bureaucratic positions (for which we expect some discretion to be present in every bureaucracy). Employment decisions in the middle tiers of the bureaucracy and for front-service providers are also significantly affected by political favouritism.
Second, when we measure the amount of support provided by a supporter to her party (namely, the amount of money donated or the number of votes brought to the party), we find that, in line with the quid pro quo nature of patronage relationships, the extent of favouritism increases relative to the amount of support provided.
Third, we find that loyalty to a party seems to be a second-order determinant of this preferential treatment. Individuals who had long been loyal to the winning party were not rewarded significantly more than those who had recently switched their loyalty. To the extent that party loyalty is a proxy for the degree of ideological alignment individuals feel to the mission of the party, this seems inconsistent with the idea that politicians are motivated primarily by the desire to increase the ideological alignment between the public workforce and a political power. In other words, the amount of support provided to the party, rather than the degree of ideological alignment to its mission, seems to be the primary driver of this favouritism.
In the second step of our analysis, we gauge the competence of the supporters more likely to benefit from their political connections. Using a host of measures of competence, we find that political connections are an especially helpful means to obtain a public job for the least competent supporters. Put differently, the provision of political support substitutes individual competence as determinant of employment decisions, consistent with patronage being an important mechanism behind our results.
Finally, our findings that discretion is used as a patronage tool suggest an additional negative implication for the effective functioning of the public sector. As documented in the context of Brazil’s local elections (Akhtari et al. 2017), significant disruption in public personnel can lead to a negative impact on the provision of public goods. Specifically, when the party in power in a municipality changes, the municipality experiences a decrease in the quality of education provision, together with a spike in turnover of local teachers and school headmasters. In light of our study, we can rationalise these spikes in turnover with politicians’ need to reward their own supporters while firing the supporters of the previous party in power. In Figure 2, we formally show the link between a municipality-specific measure of patronage (on the horizontal axis) and the degree of public sector turnover in the year following an electoral defeat of the incumbent party in that municipality (on the vertical axis), with hiring in green and job separations in red. The data show a strong positive relationship between patronage and public sector turnover: municipalities where we see larger public personnel disruptions are indeed those where we measure a higher presence of patronage practices.
Of course, our study cannot pinpoint all the ways in which political discretion in public employment decisions may affect societal welfare. For instance, individuals directly hired by politicians through discretionary means could be incentivized to exert more effort on their job, with positive implications for the quality of public services. While additional research is needed to shed light on these questions, our study highlights the need to closely monitor the way in which discretion is used. By leaving the door open to patronage practices, discretion can negatively affect the competence of the public workforce, as well as spur frequent and inefficient disruptions in a bureaucracy.
While we cannot speak to the extent to which our findings extend beyond this specific case, Brazil is considered a primary example of a professionalized and meritocratic civil service among Latin American countries (Iacoviello 2006). This suggests that the Brazilian case may well represent a lower bound for the presence of this phenomenon in the public sectors of today’s developing countries.
Akhtari, M, D Moreira and L Trucco (2017), “Political Turnover, Bureaucratic Turnover, and the Quality of Public Services”, Working Paper.
Colonnelli, E, M Prem and E Teso (2019), “Patronage and Selection in Public Sector Organizations”, CEPR Discussion Paper 13697.
Evans, A (2008), “Civil Service and Administrative Reform: Thematic Paper. Background Paper to Public Sector Reform: What Works and Why? An IEG Evaluation of World Bank Support”, Independent Evaluation Group, World Bank, Washington, DC.
Evans, P and J E Rauch (1999), “Bureaucracy and Growth: A Cross-National Analysis of the Effects of ‘Weberian’ State Structures on Economic Growth”, American Sociological Review 64: 748–765.
Grindle, M S (2012), Jobs for the Boys, Harvard University Press
Iacoviello, M (2006), “Analysis comparativo por subsistemas”, in Informe sobre la situaci´on del servicio civil en America Latina, K Echebarria, Washington, DC: Inter-American Development Bank.