摘要：The R software is commonly used in applied finance and generalized autoregressive conditionally heteroskedastic (GARCH) estimation is a staple of applied finance; many papers use R to compute GARCH estimates. While R offers three different packages that compute GARCH estimates, they are not equally accurate. We apply the FCP GARCH benchmark (Fiorentini, Calzolari and Panattoni, 1996), proposed by McCullough and Renfro (1999), which uses the Bollerslev and Ghysels (1996) daily returns data, on three R packages: fGarch, rugarch, and tseries.
其他摘要：The R software is commonly used in applied finance and generalized autoregressive conditionally heteroskedastic (GARCH) estimation is a staple of applied finance; many papers use R to compute GARCH estimates. While R offers three different packages that compute GARCH estimates, they are not equally accurate. We apply the FCP GARCH benchmark (Fiorentini, Calzolari and Panattoni, 1996), proposed by McCullough and Renfro (1999), which uses the Bollerslev and Ghysels (1996) daily returns data, on three R packages: fGarch, rugarch, and tseries.
