Economic and Ideological Motivations of Bureaucrats Turnover Decision

• In Event: Bureaucracies

Abstract

How do public officials weigh economic concerns vis-à-vis policy preferences when deciding whether to stay or quit their jobs when a change in administration is expected?

We develop a formal theory of turnover decision in the public administration in which individuals extract utility both from material and non-material features of their job. We define the material features as the pecuniary compensation that a public servant receives for performing her job. The non-material features are derived from the ideal point of the policies that the public servant helps to implement as well as from the power that she has over policy implementation (i.e. functional preferences). Given that a change in administration alters the policy ideal point and the mandate of agencies with a certain probability, the functional value (both in terms of policy ideal point and power) that public servants attach to their jobs is expected to be affected with the same probability. This probability is increasing with the difference between the ideal point of the current and expected future incumbent and with the probability of the current incumbent to leave office. Based on our theory, we put forward the following hypotheses:

• A change in government decreases the influence over policy of the average bureaucrat, thereby increasing her probability of leaving her job.
• After the change in government, bureaucrats with high policy influence have a higher probability of leaving the administration than the ones with low policy influence.
• As the deviation in agencies' mandate increases, the probability of leaving increases.
• Bureaucrats' probability of leaving is greater in agencies more severely affected by the change in administration than in less severely affected agencies.
• Before the election, as the probability of the incumbent losing the election increases, bureaucrats are more likely to leave their posts.
• Before the election, the effect of the probability of the incumbent losing the election is increasing in the expected change in the mandate of the agency.

We take advantage of employee-employer matched database of the universe of public and private employees in Brazil to test these hypotheses. We rely on labor economic literature to select labor market conditions controls that capture public officials’ economic concerns (Hall and Lazear, 1984). Furthermore, we measure deviation in agencies’ mandate based on Manifesto Project Uncertainty Estimates and on an original expert survey conducted with Brazilian academics and public officials. We also make use of polls to measure the perceived probability of a change in the administration in the months preceding presidential elections.

References
Hall, R. E. and Lazear, E. P. (1984). The excess sensitivity of layoffs and quits to demand. Journal of Labor Economics, 2(2):233–257 References

Author

• Tainá Souza Pacheco, FGV