When are Local Governments and Bureaucrats Held Accountable for the Quality of Public Services? Evidence from Brazil's Education Sector

Abstract

Under what conditions are governments held accountable by voters for the quality of public services? When voters do exercise accountability pressures, to what extent do elected officials pass them on to the bureaucracy? I address these questions by studying Brazil’s education sector, whose its institutional design is particularly favorable to local government accountability and to causal identification. Using a regression discontinuity design, I find that meeting the municipality’s pre-defined and highly visible target of school quality increases the incumbent party’s expected probability of re-election by about 12 percentage points. Bureaucrats (school directors) are generally not affected by whether they meet their target or not, regardless of whether they were elected by the community or appointed by politicians. However, when a new municipal government is elected and many school directors are replaced, appointed directors of schools that had met their performance target have a probability of turnover 7 points lower than directors of schools that had missed their target. This is consistent with incoming governments anticipating an electoral benefit from efficient directors and thus choosing which of the existing directors to replace based partly on bureaucratic effectiveness. By analyzing in a single empirical setting how the careers of politicians and bureaucrats are affected by whether they meet performance targets, the paper presents evidence of the so-called “long route to accountability.” These findings are relevant for the literatures on retrospective voting, accountability, and bureaucracies.