Functional‐coefficient regression models with GARCH errors

Abstract

The GARCH models are widely used to model various financial data with nonlinearity and heteroscedasticity structures. In this article, we propose a functional‐coefficient regression model with GARCH(r, s) errors to model these kinds of data. To deal with the effect of heteroscedasticity, we introduce a two‐step approach to estimating the unknown coefficient functions and the volatility, which results in unweighted and weighted local linear estimators. Asymptotic properties of the proposed estimators are established. Our results demonstrate that the weighted estimator is more efficient than the unweighted one, and the functional coefficients can be estimated by the weighted estimator as if the volatility was known. Both simulations and real data examples support our theoretical results. In particular, when there are GARCH effects, our two‐step estimator mimics the oracle estimator, with the true volatility being known in advance.