Asymptotic distribution of least squares estimators for linear models with dependent errors: regular designs (Q2633514)

This paper considers the linear regression model, \(Y=X\beta+\varepsilon\), with a fixed design matrix \(X\) and a stationary short-range dependent error process \(\varepsilon\). By using Hannan's central limit theorem for the least squares estimator \(\hat{\beta}\) under a mild condition on the error process and a regular design matrix, the asymptotic covariance of \(\hat{\beta}\) is characterized by the autocovariance function of the error process. Then the authors propose an estimator of the spectral density function of the error process and prove its consistency under Hannan's condition. These results can be applied to most of short-range dependent processes, some of which are also specifically exemplified in the paper.