Civil Service and Patronage in Bureaucracies∗

We develop a model of government personnel policy with electoral competition in an effort to understand when high quality bureaucracies will be created and sustained. In the model, two parties compete for office over an infinite horizon, and politicians in office choose a mix between civil servants (who produce public goods in a good bureaucracy) and patronage appointees (who produce private goods and can influence re-election). Civil servants make future good bureaucracies more likely, creating a collective action problem where each party would prefer not to pay the initial cost of investing in civil service. The analysis suggests that investment in civil service by the incumbent party is affected by the characteristics of the opposition, and by party system polarization. The electoral context is also important, and the model calls into question previous arguments that have linked electoral vulnerability to the creation of civil service. Finally, numeric results on long-term bureaucratic quality suggest that the factors affecting incentives to invest in civil service may differ in competitive democracies from factors that sustain civil service in the long-run. ∗We thank Anna Bassi, panel participants at the 2015 APSA meetings, and graduate students in Columbia’s political economy workshop for helpful comments.