This is the fourth in this year’s series of job market posts.
As the quality of a country's public sector workforce is an essential factor in its effectiveness in providing services, governments should try to hire qualified individuals in the public sector (Finan et al. 2017). However, there may be an important obstacle to the quality of this recruitment process, especially in developing countries: politicians could engage in patronage - the use of public sector jobs to reward their political supporters.
Virtually all modern bureaucracies are characterized by a civil service system, where the introduction of meritocratic hiring criteria was meant to shield public sector jobs from patronage practices. However, politicians typically retain some discretion in the selection of public workers, for instance through the use of temporary contracts (Grindle 2012). As a consequence, patronage could still play an important role in public sector hiring. Despite the potentially significant impact of this phenomenon on the quality of the public workforce, no study has systematically documented its presence in a modern bureaucracy, limiting our understanding of its consequences for the public sector.
In my job market paper "Patronage in the Allocation of Public Sector Jobs" (joint work with Emanuele Colonnelli and Mounu Prem), I study patronage in the context of Brazilian local governments during the 1997-2014 period.
A Unique Dataset On the Careers and Political Support of Brazilian Public Sector Workers
A crucial empirical challenge has made it difficult to conduct systematic investigations of patronage in public employment: we typically lack data on the careers of public sector workers and on their connections with the political power. To overcome this challenge, we combine two data sources: a longitudinal dataset with detailed information on the careers of the universe of Brazilian public sector employees between 1997 and 2014, and a dataset on more than 2,000,000 political supporters of Brazilian local parties over this period. These supporters are either local politicians or campaign donors, and based on their party affiliation or on the recipient of their donations we can clearly link them to the local party that they support.
In the first step of the analysis, we test whether political supporters of the party in power in a municipality are favored in their access to public sector jobs. To do so, we use a regression discontinuity design. Specifically, we compare the public sector careers of supporters of a municipality's elected mayor with those of supporters of the losing mayoral candidate, focusing on municipalities where the margin of victory of the winning party over the runner-up was very small. We find that:
- being a political supporter of the party in power increases the probability of being employed in the public sector by 12.4 percentage points (or 51%) among local politicians and by 6.7 percentage points (or 33%) among campaign donors (See Figure).
- Importantly, political favoritism in public hiring does not merely affect the careers of high level bureaucrats, but it is a widespread phenomenon across the whole Brazilian public sector hierarchy: we find that political supporters are granted significant preferential access to jobs from the top to the middle tiers of the bureaucracy, and also to jobs as front-service providers.
Patronage Explains Favoritism in Public Employment, With Negative Impact on Selection
While multiple mechanisms can potentially explain this result, we use detailed information on supporters' and jobs' characteristics to show that patronage is the crucial channel behind this political favoritism: these individuals are rewarded for their support, and the provision of political support substitutes an individual's quality as a hiring criterion.
First, consistent with a quid pro quo mechanism, we find that the extent of preferential access to public jobs and the associated monetary returns are monotonically increasing in an individual's amount of support provided to the party in power (measured by the amount of money donated or by the number of votes brought to the party).
Second, we find that supporters of the party in power are screened less on qualifications. Political favoritism in accessing public jobs is particularly concentrated among supporters who are unqualified in terms of education for the job in which they are hired. Among public sector workers, those who were supporters of the party in power are on average 17% more likely to be unqualified in terms of education than those who were supporting the runner-up party in the election. In addition, political favoritism is particularly strong among supporters with lower skills valued by the private sector, as measured by their previous private sector wages.
Third, we provide evidence that is inconsistent with alternative interpretations of this political favoritism in public sector hiring. Indeed, political parties do not seem to be screening their supporters on hard-to-observe dimensions of ability. Similarly, we find that supporters who recently switched political alliances are as likely to be favored in accessing a public sector job as individuals who have been loyal to the party for a long period of time. This suggests that political favoritism does not stem from an attempt to increase the ideological alignment between public workers and the party in power.
While we find that patronage negatively affects the quality of the selected public employees, there may still be positive effects of patronage which we are unable to investigate. For instance, loyalty between supporters and parties could limit agency problems and increase supporters’ effort on the job. How does the presence of patronage affect the quality of public goods provision? We provide suggestive evidence that patronage is associated with a worse provision of public services by linking differential changes in within-municipality variation in the extent of patronage over time to the quality of primary education provided by Brazilian municipalities (one of the main responsibilities of Brazilian local governments). We show that an increase in the extent of patronage in a municipality is associated with lower students’ test scores in the local public schools.
Implications of the Study
Our paper documents the important role that patronage still plays in public sector hiring. Importantly, since the Brazilian public sector is considered one of the most meritocratic in Latin America (Grindle 2010), we are probably providing a lower bound estimate of this phenomenon in the bureaucracies of modern developing countries. Of course, granting some hiring discretion to politicians can be beneficial, for instance by improving agency relationships between politicians and some layers of the bureaucracy. However, our results indicate that the patronage practices that are facilitated by this discretionary powers lead to a significant negative cost: the provision of political support substitutes for the qualifications of potential hires, lowering the average quality of the public workforce.
Edoardo Teso is a PhD student at Harvard University. More information on his research can be found on his website.