A0521
Title: Consistent autoregressive spectral estimates under GARCH-type noises
Authors: Ching-Kang Ing - National Tsing Hua University (Taiwan)
Wen-Jen Tsay - Academia Sinica (Taiwan)
Hsin-Chieh Wong - National Taipei University (Taiwan) [presenting]
Abstract: The semiparametric estimation of the spectral density function of a stationary time series driven by a general class of noise with conditional heteroskedasticity is considered. First, it is shown that the ordinary least squares (OLS) based autoregressive regression method can consistently estimate the spectral density, even though the noise is no longer a restrictive independent and identically distributed (i.i.d.) process. Secondly, this promising finding is established with much less restrictive constraints than previously imposed.
