We exploit a period of proliferation of new administrative units in Brazil between 1988 and 2010, in which 1,146 municipalities were created (an increase of 35%), to investigate the short- and medium-term effects of secessions on socio-economic outcomes. We first argue that elite capture and fiscal incentives play an important role in secessions. Because the decision to secede is not random, we collect data on municipalities that had secession requests denied due to a Constitutional Amendment that curbed the formation of new municipalities after 1996 to create a control group for municipalities that seceded. Using past tract-level Census data to reconstruct outcomes for new boundaries, we find that secession is associated with better education, health, wealth, and public service outcomes. We document that the positive effects are mostly driven by new municipalities, while old municipalities present negligible changes. We show that increases in revenues do not fully explain our findings and we discuss further mechanisms, such as changes in state capacity, infrastructure, and migration.
Decentralized assignments in the education market have been increasingly replaced by centralized ones. However, empirical evidence on these transitions is scarce. This paper examines the adoption of centralized admissions in the Brazilian higher education market. Using rich administrative data, we exploit time variation in the adoption of a clearinghouse across public institutions to investigate the impacts on student migration, sorting, and enrollment. We find that institutions under the centralized assignment are able to attract students from other locations and with substantially higher test scores. We find no effects on enrollment rates. In addition, we document lower migration rates for students enrolled in institutions that could not switch to centralized admissions, particularly the private ones, indicating that the market becomes more local for them.