frontier (version 1.1-8)

coef.frontier: coef method for class frontier

Description

Extract the coefficients from stochastic frontier models returned by frontier.

Usage

# S3 method for frontier
coef( object, which = "mle", extraPar = FALSE, … )

Arguments

object

an object of class frontier (returned by the function frontier).

which

character string. Which coefficients should be returned? ('start' for starting values provided by the user, 'ols' for coefficients estimated by OLS, 'grid' for coefficients obtained by the grid search, or 'mle' for coefficients estimated by Maximum Likelihood).

extraPar

logical. If TRUE, additional parameters are returned: sigmaSqU = sigmaSq * gamma (with $u$ ~ $N^{+}$( mu, sigmaSqU )), sigmaSqV = sigmaSq * ( 1 - gamma ) (with $v$ ~ N( 0, sigmaSqV )), sigma = sigmaSq^0.5, sigmaU = sigmaSqU^0.5, sigmaV = sigmaSqV^0.5, lambdaSq = sigmaSqU / sigmaSqV, and lambda = sigmaU / sigmaV. Please note that sigmaSqU and sigmaU are not the variance and standard error, respectively, of $u$. If the model is an error components frontier and argument timeEffect is FALSE, also the following additional parameters are returned: varU = the variance of $u$, sdU = varU^0.5, and gammaVar = varU / ( varU + sigmaSqV ). Please note that the variance of $u$ usually differs between observations if the model is an error component frontier with ‘time effect’ or an efficiency effects frontier.

…

currently unused.

Value

coef.frontier returns a named vector of the coefficients.

See Also

coef.summary.frontier and sfa.

Examples

# NOT RUN {
   # example included in FRONTIER 4.1
   data( front41Data )

   sfaResult <- sfa( log( output ) ~ log( capital ) + log( labour ),
      data = front41Data )
   coef( sfaResult, which = "ols" )
   coef( sfaResult, which = "grid" )
   coef( sfaResult )
# }