Using dynamic compositional models to study public opinion and taxation in Brazil

• Grant number: 17/11559-0
• Support Opportunities: Scholarships abroad - Research Internship - Doctorate
• Start date: August 14, 2017
• End date: August 13, 2018
• Field of knowledge: Humanities - Political Science - Public Policies
• Principal Investigator: Lorena Guadalupe Barberia
• Grantee: Natália de Paula Moreira
• Supervisor: Guy David Whitten
• Host Institution: Faculdade de Filosofia, Letras e Ciências Humanas (FFLCH). Universidade de São Paulo (USP). São Paulo , SP, Brazil
• Institution abroad: Texas A&M University, United States
• Associated to the scholarship: 16/13199-8 - Preferences and Public Policy Congruence, BP.DR

Abstract

Brazil has a regressive tax system and a high concentration of low-income voters (Gobetti and Orair, 2016). Low-income voters pay proportionally more on taxes than high-income voters. In spite of the relevance of understanding how voters from different income groups evaluate changes on specific taxes, there are no studies assessing whether governments are responsive to voters on this issue for Brazil. I intend to cover this gap. In this sense, I will analyze how voters evaluate changes in tax policy in Brazil, and whether voters from different economic groups evaluate tax policy differently depending on the type of tax. I will employ a measure that captures trade-offs between how different types of voters (high, middle and low-income) evaluate the president. To do so, I will create a compositional dependent variable that captures the changes in the share of voters of each category that approve of the president. To analyze this dependent variable in a time series setting, I will employ a dynamic pie approach that is used for modeling trade-offs in compositional variables over time (Philips, 2016; Philips et al., 2016). (AU)