Variance-Covariance Matrix for a Fitted Quantile Ratio Regression Model Object

This functions returns the variance-covariance matrix of the coefficients of a fitted qrr model object.

Usage

# S3 method for class 'qrr'
vcov(object, method = "approximate", R = 200, update = TRUE, ...)

object: an object of class qrr.
method: if "approximate", the variance-covariance estimate is approximated by the inverse of the numerical Hessian. The latter is calculated as detailed in Farcomeni and Geraci (2023). If "boot", the variance-covariance estimate is calculated by means of ordinary bootstrap (see boot).
R: the number of bootstrap replications.
update: logical flag. If TRUE (the default), the statistic to be resampled is obtained via an update of the qrr object. If FALSE, then the statistic to be resampled is obtained via a do.call of the qrr object. See details.
...: not used.

The use of update = FALSE is preferred when the function vcov.qrr is called from within another function.

Farcomeni A. and Geraci M. Quantile ratio regression. 2023. Working Paper.

Marco Geraci with contributions from Alessio Farcomeni