Semi-parametric estimation for ARCH models

In this paper, we conduct semi-parametric estimation for autoregressive conditional heteroscedasticity (ARCH) model with Quasi likelihood (QL) and Asymptotic Quasi-likelihood (AQL) estimation methods. The QL approach relaxes the distributional assumptions of ARCH processes. The AQL technique is obtained from the QL method when the process conditional variance is unknown. We present an application of the methods to a daily exchange rate series. Keywords: ARCH model, Quasi likelihood (QL), Asymptotic Quasi-likelihood (AQL), Martingale difference, Kernel estimato