Occupy Government: Democracy and the Dynamics of Personnel Decisions and Public Sector Performance

We study the causes and consequences of patronage in Brazilian cities since the country’s re-democratization. We test key mechanisms – fiscal rules, accountability, political ideology, and rent-seeking – and estimate the consequences of patronage for public finances. Our data consist of the universe of public sector employees merged with their party affiliations, and a dynamic regression discontinuity design is applied to disentangle patronage from the growing political participation. The short-term patronage effect is large, with winning political coalitions increasing their shares of public sector workers and wages by 4 and 6 percentage points, respectively, during a mayoral term. Part of this effect lasts longer than a decade, with winning coalitions also occupying civil servant jobs that perform service-oriented tasks. This political occupation of government jobs is not associated with ideology, though. Instead, lack of accountability and rent-seeking are the primary driving forces, while reliance on intergovernmental transfers only increases patronage for smaller cities. Higher patronage does not affect the size of local governments, but it changes the composition of expenditures: hiring politically connected workers crowds out, almost one-to-one, non-affiliated employees. Overall, patronage accounted for more than half of the dramatic increase in public sector political employment since the Brazilian re-democratization.