Essays on macroeconometrics and financial econometrics: a Bayesian approach

• Autor(a) principal: Alves, Cássio Roberto de Andrade
• Data de Publicação: 2023
• 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/96/96131/tde-28082023-141853/
• topic: Bayesian estimation; Econometria bayesiana; Finanças; Função de resposta ao impulso; Inflation persistence; Instrumental variable; Macroeconomia; Monetary policy; Previsão; Sentiment analysis; Shrinkage prior; Stochastic volatility; Variáveis instrumentais; Yield curve

The goal of this doctoral dissertation is to showcase the importance and applicability of Bayesian econometrics in the realm of financial and macroeconomic analysis. The thesis is formed by four independent essays in macroeconometrics and financial econometrics, in which the Bayesian estimation is the common factor of the four essays. In the first article, a heteroscedastic structure with jumps is added to a structural model to measure inflationary persistence in Brazil from 1995 to 2019. The general result is that including stochastic volatility with jumps reduces intrinsic inflation persistence. In the second article, the Brazilian Central Bank Communication is used to predict the yield curve. The results indicate that the Central Bank Communication, measured by the sentiment of the Central Bank, helps to predict the term structure of interest rate. In the third article we explore the the analysis of Central Bank Communication and yield curve in a in-sample perspective. The objective of this article is to investigate the bidirectional relation between the Brazilian Central Bank Communication and the yield curve. The results shown that the Central Bank Communication can shape yield curve curvature and slope, and that there is a strong relation between monetary authority communication and the curvature of the yield curve. Both article two and three present evidence that words of monetary authority impacts market players, making it a valuable instrument for monetary policy. The last article proposes an instrumental variable Bayesian shrinkage approach to estimate the Capital Asset Pricing Model (CAPM) using a large set of instruments. Using simulated data, the proposed approach reduces the bias of estimation, caused by measurement errors. In an empirical application, the proposed method delivered a different estimation from the traditional approach and this difference increases the explanatory power of the Capital Asset Pricing Model in explaining the variation in the average cross-sectional returns of assets.