Proteção cambial com contratos futuros: uma análise comparativa da efetividade de modelos de hedge

• Ano de defesa: 2005
• Autor(a) principal: Daniel Loureiro Araujo
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; UFMG
• Palavras-chave em Português: correção de erros; Proteção cambial; Volatilidade condicional; Cointegração; GARCH bivariado; Efetividade de hedge; Modelo de; Câmbio; Administração; Hedging (finanças)
• Link de acesso: http://hdl.handle.net/1843/BUBD-99XHWE

Since the Brazilian government has implemented the floating currency regime in the early 1999, the exchange rate risk has become an important issue for the companies financial management. The aim of this research was to assess the hedge effectiveness of some of thehedging models, in order to find out which of them would be the best choice for exchange rate risk management. Comparing with the no hedge situation, four models were implemented and analized in terms of variance reduction: the naïve model, which assumes that the hedge ratiomust always be equal to one for the whole period; the conventional model, estimated by ordinary least squares and the error correction model, with and without a GARCH error structure. As suggested by Lien, Tse and Tsui (2002), the conventional model and the error correction model were estimated in a dynamic way, through a rolling window procedure, in order to make the comparison with GARCH models more fair. As a general result, the error correction model has shown to be superior compared to the others, although this superiority over the conventional one was just marginal. Moreover, the use of the nearby futures contract has shown to be more effective in the risk reduction compared to the futures contracts series that has been rolled over 28 days before the expiration date. There was no clear evidence ofthe influence of several aspects on the hedge effectiveness, such as the size of the estimation window, the readjustment frequency of the hedge ratios, and the volume of currency exposition. Concerning the division of the portfolio returns in three subperiods, the results were quite mixed. The GARCH models seemed to be more effective in the risk reduction only in the last subperiod, although this superiority was only marginal. One can conclude that the dynamic estimation of the conventional and error correction models has been enough to take account of the time varying feature of the hedge ratios. The GARCH model results were notsuperior enough to compensate their implementation and estimation costs.