Agent-based modeling for understanding the interactions among federal, state and municipality government on the distribution of gains in educational and health policies

• Grant number: 15/14240-9
• Support Opportunities: Scholarships in Brazil - Post-Doctoral
• Start date: August 01, 2015
• End date: July 31, 2016
• Field of knowledge: Applied Social Sciences - Economics - Social Welfare Economics
• Principal Investigator: Marta Teresa da Silva Arretche
• Grantee: Davoud Taghawi-Nejad
• Host Institution: Centro Brasileiro de Análise e Planejamento (CEBRAP). São Paulo , SP, Brazil
• Associated research grant: 13/07616-7 - CEM - Center for Metropolitan Studies, AP.CEPID

Abstract

Brazilian Educational and Healthcare Funding Structure, an ABM An agent-based model to understand the effect of interactions between federal, state and local governance on the distribution of educational and health policy outcomes. Income inequality has declined steadily in Brazil, and much less is known about the equality of service provision. In the healthcare and educational fields, the answers to these problems are not straightforward. The Brazilian tax spending system is a complex network, where interactions between spending requirements, funding sources and transfers make it hard to measure the distribution of service provision. In addition, they make it hard to predict the outcome of policy measures on the equality of service provision. For example every municipality must spends at least 25% of its revenue on education. However, a municipality must spend a set amount of money per student. Municipalities whose 25% requirement is not sufficient to comply with the per-student requirement get a transfer from other municipalities who exceed the second requirement. If the other municipalities' resources are exhausted money comes from the state. This is supplemented by direct payments from the state and the federation. In short a detailed modeling of the funding structure is necessary to judge whether the federal system inhibits or fosters inequality. To test policies, we are creating an ABM that simulates the education and healthcare spending on the union, state and municipal levels. The model has 5,597 agents: the union, 26 Brazilian states and 5,570 municipalities. Each of the municipal agents is characterized by its educational and healthcare outcomes. This will allow us to predict the distributional effect of policies and spending on the municipalities. In addition, as we model the effectiveness of education and healthcare spending, we can make predictions about the distributional effects in educational and healthcare outcomes. The union agent has a certain tax revenue that it can spend on education and healthcare; the states and municipalities are similar. All agents have decision rules and legal constraints that establish the allocation of their taxes. However, the researcher or policymaker can also override the decision rule to make predictions about a policy he or she wants to test. Based on their tax revenue, the decision rule and legal constraints, all agents spend money on education and healthcare programs. The federal state spends money on municipalities and transfers money to states. The states spend money on municipalities. Municipalities spend money on themselves and transfer money to other municipalities when they are obliged to do so. The decision rules that mimic real-world spending patterns. The constraints include constitutional and legal minimum spending rules. We can now test the different spending rules; different revenue schemes and a different federal structure, such as centralization or decentralization. This will give us insights about distribution of municipal spending on education and health; distribution of educational and health outcomes; evolution of the total educational and health spending for individual municipalities, states and the federation; evolution of educational and health outcome for individual municipalities and in aggregate For example, we could investigate the distributional outcome if the states do not spend on municipalities and all spending comes from the federal government. Further, we could test what would happen if no redistribution occurred or if all spending was centralized. (AU)