Estudo das relações de paridade de juros para a economia brasileira no período recente

• Ano de defesa: 2002
• Autor(a) principal: Takami, Marcelo Yoshio
• Orientador(a): Leme, Maria Carolina da Silva
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Palavras-chave em Português: Paridade; Risco; Prêmio; Cointegração
• Palavras-chave em Inglês: Exchange rate; Interest rate; Parity; Premium; Cointegration; Risk
• Link de acesso: http://hdl.handle.net/10438/1842

The aim of this paper is to test whether the correction of the interest parity conditions by market expectancy (uncovered parity) and by risk premium (covered and uncovered parity) results in a (close) econometric validation of the non-arbitrage relationship. In order to accomplish this goal, we combined domestic and non-domestic fixed rate instruments and applied time series econometrics. To start with, we tested the traditional interest parity condition (covered and uncovered). In the next step, we tested again but with risk premium incorporated. In the case of UIP (uncovered interest parity), we didn't obtain satisfactory results, even adjusting for risk premium. This adjustment led to correct coefficient signals, but the magnitude of the exchange devaluation coefficient got too high. Even though we had obtained CIP (covered interest parity) validity, this result was not expected, as this would imply that the country risk premium was null along this period. After adjusting the CIP for default risk premium, the series don't cointegrate any longer, i.e., the default risk premium would have a behavior independent from future premium and interest differential. The possible reasons for the non-expected results are: sample intervalless than 3 years, data measurement error or simuItaneous control of exchange rate and interest rate by monetary policy authorities.