Date of Award

January 2015

Degree Type


Degree Name

Doctor of Philosophy (PhD)


Social Sciences


Peter Wilcoxen


Brazil, fiscal policy, local government, political economy, political institutions, public budgeting

Subject Categories

Social and Behavioral Sciences


This dissertation covers different aspects of political institutions and fiscal policy in local governments. The first essay examines the political consequences of an increase in the district magnitude (the number of open seats) for legislatures. The second essay studies how the size of a local legislature affects the size and composition of public spending. The third essay develops and tests a theory of how transaction costs might affect the use of earmarked revenues by local budgeting authorities. The last essay compares the relative impacts of factors that might make a local government more likely to adopt own-source revenues. These questions are examined in the context of Brazilian and US local governments. Individual chapter abstracts follow.

Chapter 1: Research has consistently identified district magnitude as an important institutional factor in political competition. The most familiar of these formulations is Duverger’s hypothesis, which argues that district magnitude affects the number of parties and legislative fragmentation among parties. Yet parties play little to no role in structuring political competition in some polities. How might Duverger’s rule apply in the absence of meaningful parties? Using the case of Brazilian municipalities, this paper estimates the influence of changes in district magnitude on alternative conceptions of political competition – the number of personalist political coalitions and opposition strength. Results indicate that large changes in district magnitude had a positive but minimal effect on the number of coalitions in government. The most consequential result of a larger district magnitude was a strengthened opposition, which increased its numbers. Evidence from fieldwork suggests that larger city councils in Brazil may be more autonomous from the executive and engender a better check on executive power.

Chapter 2. Many factors contribute to the size and composition of public spending. Research that examines the role of legislature size has traditionally focused on political systems where the legislature has budgeting authority, and higher expenditures emerge as a result of logrolling. This paper develops an alternate view of legislature size and expenditures for a political system where an independent executive holds the budgeting authority and must spend fiscal resources to maintain his or her legislative coalition. I examine the theory in the context of Brazilian municipalities. Results indicate that larger legislatures led to higher public spending. An average municipality that expanded from 10 to 15 legislators in 2013 spent 3% to 4% more than a similarly situated municipality that maintained 10 legislators. This equals R$400k per year, or 65 minimum wage jobs, per additional legislator. There was an appreciable increase in urban infrastructure spending, while education expenditures did not change systematically in response to larger local councils. The overall increase and the pattern of the increase are consistent with the notion that Brazilian mayors use public spending to build and maintain their legislative coalitions, and that they must appropriate more public funds to achieve these goals when the legislature is larger.

Chapter 3. The question of whether earmarked public revenues are spent on their intended purpose has a long history in the academic literature. Theory suggests that earmarks should not matter since budget authorities can always shuffle money between accounts. However, empirical studies often find that much of marginal earmarked revenue is spent overwhelmingly on its intended purpose. There are many theories for this outcome, and this paper adds one more to the literature: transaction costs in the budgeting process. When transaction costs are low, governments optimize between expenditure categories. When the costs are high, the earmarked funds are spent exclusively on their intended purpose. Further development of this theory suggests governments that are highly reliant on earmarked revenues will face unnecessary funding shortfalls in some spending categories, and will be more likely to pay the costs of re-allocating the revenues. The theory is tested empirically using the case of Oklahoma counties and their management of the county highway fund from 1973 to 2012. Results are in line with the theory, suggesting that a moderate amount of earmarked revenues stimulate specific expenditures, but the effect diminishes as earmarks become a larger portion of the government’s total revenues.

Chapter 4. There is increasing recognition among scholars that the fiscal link between citizens and their governments is an important contributor to more efficient allocation of public resources and a more accountable government. However, many local governments in developing countries receive the majority of their funding from intergovernmental transfers, which severs that fiscal bond. There are some cases in which local governments, despite large intergovernmental transfers, have chosen to increase their own taxes. Understanding why local governments choose to adopt new own-source revenues is an important step in the search for policies that encourage more local fiscal effort. This paper examines this question in the context of a local lighting tax in Brazilian municipalities. Results indicate that Brazilian municipal governments are willing to engage in the politically difficult work of taxing their citizens in times of revenue shortfall and service need, but that political calculations and the influence of nearby municipalities play a bigger role in the decision to adopt a local tax.


Open Access