Estimation of a log-variance model

Two-step estimation of a log-variance model: OLS in step 1, bias correction w/residuals in step 2 (see the code for details). The function larchEstfun() is not intended for the average user, but is called by larch and gets.larch.

Usage

larchEstfun(loge2, x, e, vcov.type = c("robust", "hac"), tol = 1e-07)

Arguments

loge2

numeric vector, the log of the squared errors 'e' (adjusted for zeros on e, if any)

x

numeric matrix, the regressors

e

numeric vector, the errors

vcov.type

character vector, "robust" (default) or "hac". If "robust", then the White (1980) heteroscedasticity-robust variance-covariance matrix is used for inference. If "hac", then the Newey and West (1987) heteroscedasticity and autocorrelation-robust matrix is used

tol

numeric value. The tolerance for detecting linear dependencies in the columns of the regressors in the first step estimation by OLS, see ols. Only used if LAPACK is FALSE

Details

No details for the moment.

Value

A list.

References

No references for the moment.

Author

Genaro Sucarrat, http://www.sucarrat.net/

See also

qr, larch, gets.larch

Examples

##no examples for the moment