vcov.mnprobit {switchSelection}

Calculate Variance-Covariance Matrix for a mnprobit Object.

Description

Return the variance-covariance matrix of the parameters of mnprobit model.

Usage

## S3 method for class 'mnprobit'
vcov(
  object,
  ...,
  type = object$cov_type,
  regime = NULL,
  n_cores = object$other$n_cores,
  n_sim = object$other$n_sim
)

Arguments

object

an object of class mnprobit.

...

further arguments (currently ignored).

type

character representing the type of the asymptotic covariance matrix estimator. If

regime

non-negative integer representing the regime of the two-step procedure for which covariance matrix should be returned. If estimator = "2step" and regime = NULL then covariance matrix of the first step parameters' estimator will be returned. Otherwise the function estimates covariance matrix for the second step parameters associated with corresponding regime.

n_cores

positive integer representing the number of CPU cores used for parallel computing. If possible it is highly recommend to set it equal to the number of available physical cores especially when the system of ordered equations has 2 or 3 equations. estimator argument of mnprobit is "ml" then type may be changed to any value available for input argument cov_type of mnprobit. Otherwise type will coincide with cov_type output value of mnprobit.

n_sim

integer representing the number of GHK draws when there are more then 3 ordered equations. Otherwise alternative (much more efficient) algorithms will be used to calculate multivariate normal probabilities.

Details

Argument type is closely related to the argument cov_type of mnprobit function. See 'Details' and 'Usage' sections of mnprobit for more information on cov_type argument.

Value

Returns numeric matrix which represents estimate of the asymptotic covariance matrix of model's parameters.