This function gives the estimation of the asymptotic covariance matrix of the normalized least squares estimator in the case of the linear regression model with strictly stationary errors.
cov_matrix_estimator(object)an object of class slm.
This function returns the estimation of the asymptotic covariance matrix of the normalized least squares estimator.
The function computes the covariance matrix estimator of the normalized least squares estimator from the vector cov_st
of a slm object. If the user has given the argument Cov_ST in the slm object, then it is used
to compute the final covariance matrix. If the method used is the "hac" method, then the final covariance matrix is computed via
the kernHAC function of the sandwich package, by using the Quadratic Spectral kernel and the bandwidth described in Andrews (1991).
For the methods "efromovich", "kernel" and "select", the covariance matrix estimator may not be positive definite. Then we apply the
"Positive definite projection" algorithm, which consists in replacing all eigenvalues lower or equal to zero with the smallest
positive eigenvalue of the covariance matrix.
D. Andrews (1991). Heteroskedasticity and autocorrelation consistent covariant matrix estimation. Econometrica, 59(3), 817-858.
E. Caron, J. Dedecker and B. Michel (2019). Linear regression with stationary errors: the R package slm. arXiv preprint arXiv:1906.06583. https://arxiv.org/abs/1906.06583.
A. Zeileis (2004). Econometric computing with HC and HAC covariance matrix estimators.
The R package sandwich.
kernHAC for HAC methods.