Revisiting the Politicians’ Dilemma: Electoral Cycles, Partisanship and Bureaucratic Turnover

Politicians face a dilemma when selecting bureaucrats, because they face a tension between hiring political loyalists and having a technically competent bureaucratic workforce. The study of this tension in a systematic way has been hindered by the lack of individual and longitudinal data on public employees. In this paper, I use a novel panel-structure dataset that includes the work trajectories of nearly all employees of the Chilean central government (excluding the health sector) between 2006 and 2020 to study bureaucratic turnover after (and between) presidential elections. I further match employees’ information with party membership data and university admissions test scores (a measure of ability or merit) for a large subsample of bureaucrats, to understand if these characteristics affect turnover. I first use survival models to show that – unlike in more developed civil services – there is a sizable and significant “post-election employment shock” to the bureaucracy, though it is much larger at the top and upper-middle levels of the bureaucratic hierarchy. Party membership is also a significant risk factor at these levels. I then turn to an agency-level analysis to show that there is a broad spectrum of politicization between agencies, ranging from the fully autonomous to the ones that seem to be political spoils given to low-merit, partisan bureaucrats. Indeed, I show that agencies with higher party membership have on average less competent professional and managerial personnel, and that this effect is not driven by the partisans themselves – rather, more politicised agencies appear to attract less competent (non-partisan) professionals. Overall, this study is a proof of concept of the way in which individual-level data can be leveraged to study bureaucratic politicization.