CFE 2015: Start Registration
View Submission - CFE
A0548
Title: On the need of estimating the parameters of GARCH models Authors:  Cindy Shin Huei Wang - HSBC Business School, Peking University (China) [presenting]
Abstract: A new estimating tool is proposed for a regression model with serial correlation and generalized autoregressive conditional heteroscedasticity (GARCH). This tool is constructed on the basis of the least absolute deviation (LAD) and the autoregressive (AR) approximation,namely the COAR-LAD estimation. The estimation is easy to implement and avoids issues commonly found in time series literature. Neither does the COAR-LAD estimation need to estimate many parameters of the GARCH models nor does it need to tackle specification problems one at a time. We further show that a convergent usual $t$-statistic based on our new estimator can be constructed for previously analyzed spurious regression cases, even though the exact form of the error term is unknown in practice. The simulation results indicate that the finite sample performance of our methodology is promising even in sample sizes. This finding sheds some new light on the nature of a regression with serial correlation and GARCH effect simultaneously.