Peer effects in consumption: a spatial demand system estimation using data from Brazil

• Autor(a) principal: Li, Denise Leyi
• Data de Publicação: 2021
• Tipo de documento: Tese
• Idioma: eng
• Título da fonte: Biblioteca Digital de Teses e Dissertações da USP
• Texto Completo: https://www.teses.usp.br/teses/disponiveis/12/12138/tde-23072021-195120/
• topic: Consumo; Consumption; Demand system; Efeitos de pares; Elasticidade da demanda; Elasticity of demand; Peer effects; Sistema de demanda

Household consumption is traditionally modeled without considering the influence of peers. Studying peer effects in consumption is relevant, as its presence can affect the intratemporal and intertemporal consumption decisions of households leading to distortions in social welfare. Additionally, there may be impacts on the planning and evaluation of policies involving consumption, such as income transfer programs and tax policies. The literature of social effects is theoretically and empirically challenging, with data availability being one of the main constraints. Few studies focus on disaggregated consumption, and almost none take a further step to analyze the resulting elasticities. Given that the spillover and feedback effects are relevant, price and income effects on demand can now be influenced by the social interactions between households. This thesis\'s main contribution is twofold: it develops a theoretically and empirically well-specified spatial demand system to estimate peer effects in disaggregated consumption, and it provides elasticity of demand specific to direct, indirect, and total impacts. This thesis investigates the peer effects in the composition of household expenditure using disaggregated consumption data of households from the city of São Paulo, Brazil, surveyed by Fipe between 2008 and 2013. Data about the exact addresses of all households were available and used to define household\'s reference groups characterized by geographical proximity with more precision. The spatial simultaneous demand model is an adaptation of the classic Almost Ideal Demand System. In addition, a two-stage budgeting system is considered, where consumers first choose how to spend their income among broad groups of products, which include food, and in the second stage, expenditure allocated to the food group is then distributed among food subcategories. The results of the empirical investigation suggest that neighbors cause a small but not negligible effect on a household\'s consumption behaviors. Consequently, the elasticities derived from the model with social interactions differ from those calculated from the traditional demand model.