Previsão de crescimento econômico a partir da análise textual das atas do copom (2013-2023) Fortaleza 2024

• Ano de defesa: 2024
• Autor(a) principal: Sousa, Maressa Soares de
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Palavras-chave em Português: crescimento econômico; análise textual; índice de sentimento; CNPQ::CIENCIAS SOCIAIS APLICADAS::ECONOMIA
• Link de acesso: http://repositorio.ufc.br/handle/riufc/77406

This study investigates the influence of sentiments expressed in the minutes of the Monetary Policy Committee (COPOM) of the Central Bank of Brazil (BACEN) on economic growth, using textual analysis in the context of emerging information technologies. The research covers a period of 11 years, from 2013 to 2023, analyzing 88 documents to determine how recorded perceptions and expectations can predict future economic variations. The methodology employed was biphasic: initially, a prediction index was constructed based on the Loughran-McDonald Dictionary, adapted to the Brazilian context; subsequently, simple linear regression and autoregressive models of order 2 (AR(2)) were applied. This approach was crucial in a global context of uncertainties, where traditional methods often fail to capture all the nuances that affect economic trends. The results showed a significant correlation between the sentiments expressed in the minutes and subsequent changes in economic growth, confirming the utility of the dictionary method as a tool to improve this prediction. These insights are valuable for guiding economic policies and investment strategies, reinforcing the relevance of integrating new analytical techniques to enhance the precision and reliability of economic models.