Ensaios sobre modelos de previsão econômica

• Ano de defesa: 2018
• Autor(a) principal: Godeiro, Lucas Lúcio
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal da Paraíba; Brasil; Economia; Programa de Pós-Graduação em Economia; UFPB
• Palavras-chave em Português: Incerteza macroeconômica; Regressão de texto; Aprendizado de máquina; Previsão macroeconômica; Previsão financeira; Macroeconomic uncertainty; Text regression; Machine learning; Macroeconomic forecasting; Financial forecasting; CNPQ::CIENCIAS SOCIAIS APLICADAS::ECONOMIA
• Link de acesso: https://repositorio.ufpb.br/jspui/handle/123456789/15198

This dissertation encompasses three chapters that study Economic Forecasting Models. Below are the abstracts for each chapter. Chapter 1: Measuring Macroeconomic Uncertainty to Brazil ThechapterproposesestimatingamacroeconomicmeasureofuncertaintytoBrazil. The indexwasconstructedbasedonthemethodologyofJurado(2015)usedtobuildthesame index for the US economy. We show that an increase in the uncertainty level leads economic recessions. Moreover, the recent macroeconomic policy adopted by the Brazilian government in 2010-2011 was followed up by substantial increase in the uncertainty level of the Brazilian economy. Our results suggest that the proposed uncertainty measure can be used to assess macroeconomic policies as well as predict economic recessions.Chapter 2: Identifying the Predictive Power of FED Minutes This chapter proposes a novel method to extract the most predictive information from FED minutes. Instead of considering a dictionary (set of words) with a fixed content, we construct a dictionary whose content is allowed to change over time. Specifically, we utilize machine learning to identify the most predictive words (the most predictive content) of a given minute and use them to derive new predictors. We show that the new predictors improve forecast accuracy of Output growth by a statistically significant margin, suggesting that the combination of machine learning and text regression can be interpreted as a powerful device for out-of-sample macroeconomic forecasting. Chapter 3: Equity Premium Forecasting: Identifying the Predictive Power of Financial News This chapter proposes a novel method to extract the most predictive information from Financial News published in the Wall Street Journal and The New York Times. Instead of considering a dictionary (set of words) with a fixed content, we construct a dictionary whose content is allowed to change over time. Specifically, we utilize machine learning to identify the most predictive words(the most predictive content) of a given financial news and use them to derive new predictors. We show that the new predictors improve forecast accuracy of Equity Premium by a statistically significant margin. We also finds that the Financial News increases the utility and financial gains, for a investor with a mean-variance utility function.