Who Pays for the Welfare-State? Preferences for Redistribution and Subjective Perceptions of the Costs Thereof

Author(s)

Rodrigo Sequerra Mahlmeister

• Document type: Master's Dissertation
• Press: São Paulo.
• Institution: Universidade de São Paulo (USP). Faculdade de Filosofia, Letras e Ciências Humanas (FFLCH/SBD)
• Defense date: 2021-09-15

Examining board members

Marta Teresa da Silva Arretche; Mario Fuks; Cesar Zucco Junior

• Advisor: Marta Teresa da Silva Arretche

Abstract

To assume rational voters with full information implies that the redistribution level in a democracy mirrors the well-being maximization of each citizen, as proposed by the canonical model of Meltzer and Richard. Deriving demands from the income individuals earn, however, is at odds with the empirical regularity of the lack of polarization between the poor and the rich in their attitudes toward progressive taxation. This dissertation examines cognitive gaps that may weaken the link between income and support for welfare provision, either because individuals fail to associate policies with their potential costs, or because they are usually unable to predict if they would benefit from it or suer drawbacks. Thus, one of the variables of interest that the research focuses on is the so-called \"self-stratification\", which stands for the subjective identification with income groups. It expresses the individual perception of belonging to the category of the least or the most well-o in the country. The degree to which agents are rational when taking a stance on the welfare state cannot be understood unless this perceptive dimension is accounted for, as it aects the expected return on government action. Supposedly, assuming oneself among the poor would boost support due to the perspective of being well-o after redistribution, while identifying as rich should encourage the behavior of avoiding future losses through taxation. The dissertation uses as inputs empirical data from representative opinion polls in Brazil, including an original survey-experiment, which are analyzed using quantitative methods. The first step was to investigate the origins of the phenomenon of self-stratification; that is, the factors that explain the identification with rich and poor groups. The second step was to examine the eect of these perceptions in support for redistribution. I show that preferences and the role played by its determinants are conditional on how the costs of policies are presented in survey questionnaires. In other words, individual demand, and how it is related to actual and perceived income positions in the national ranking, depend on the information mentioning the tax increase, and, especially, who would pay for it. Voters, in general, do not associate state intervention into economic inequality with taxes or transfers, and the composition of the groups that would benefit or suer drawbacks from these policies is not given. Faced with survey questions conventionally used as proxies of support for progressive taxation, for instance, there is evidence that respondents resort to their preconceived notions to infer whether they would belong to the net contributors\' group, which may weaken the relationship between actual socioeconomic position and preference. The data also shows, through correlational and causal evidence, that the provision of information regarding redistribution costs has a substantial eect on individual preference. Providing respondents with specific monetary threshold values to define contributors tends to boost the explanatory power of the income variable. The more this cuto value divides society, the more polarized their levels of support across strata become. One of the main conclusions is that, in the context of reduced uncertainty on the eects of redistribution, individual behavior approaches the rationality assumed by Meltzer and Richard. This challenges consensus established in the literature and suggests caution before ruling out the self-interest hypothesis based on opinion data. Voters, despite supporting the increase in social spending, are averse to the risk of being taxed. In addition to the contributions oered to the field of political behavior, the research also provides critical empirical implications that call for a revision of currently popular survey methodologies and measurement tools of behavioral analyses. (AU)

• FAPESP's process: 19/03725-2 - Self-stratification bias in Brazil: the role it plays on preferences for redistribution
• Grantee: Rodrigo Sequerra Mahlmeister
• Support Opportunities: Scholarships in Brazil - Master