Análise sobre o comportamento do câmbio nominal brasileiro durante a pandemia

• Ano de defesa: 2022
• Autor(a) principal: Almeida, João Pedro Arbex Octaviano de
• Orientador(a): Muinhos, Marcelo Kfoury
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Palavras-chave em Português: Câmbio; Economia; Taxas de câmbio - Brasil; Kalman, Filtragem de; COVID-19 (Doença); Covid-19; Ordinary least square
• Palavras-chave em Inglês: Nominal foreign exchange rate; Kalman Filter OLS; Nominal foreign exchange short term forecast
• Link de acesso: https://hdl.handle.net/10438/32815

During the Covid-19 pandemic, several financial assets had their prices changed abruptly. This was mainly due to a sudden change in the perception of risk of the global economy. Especially in emerging economies, this effect was exacerbated by concerns about the fiscal situation of these economies, often already fragile. Thus, we seek to understand the behavior of the nominal exchange rate of the Brazilian Real against the US Dollar during the Covid-19 crisis. With a database of 199 monthly observations since January 2006, it is concluded through the method of Ordinary Least Squares in time series, and Kalman Filter, (KALMAN, 1960), that the nominal exchange rate showed a detachment from global fundamentals mainly in 2020, the first year of the pandemic. It was noted with the help of Dummies and the Kalman Filter that the coefficients of some explanatory variables underwent significant changes from 2020 onwards, which suggests a change in the econometric relationship between the variables in line with the concepts indicated by (JR, 1976) and (WOLFF, 1987). It was also found that when comparing short-term projections using the Pseudo Out Of Sample Forecast methodology, proposed by (WOLFF, 1987), from previous periods and during the pandemic, both the structural models and the benchmarks Random Walk proposed by (MESSE; ROGOFF, 1974) and the market expectations released by the Central Bank of Brazil, lost predictive capacity. Finally, it was shown that by incorporating variables that reflect intrinsic conditions of the Brazilian economy into the model, such as the 10-year NTN-B (National Treasury Note) rate, the benchmark interest rate - SELIC, there was an improvement in short-term predictive ability during the pandemic period. We concluded that between the two, the SELIC was more relevant to this end.