SURest {panelSUR}

EC SUR System Models Estimation on (Unbalanced) Panel Data

Description

SURest is used to estimate one-way and two-way SUR systems on unbalanced panel data by GLS estimator also allowing cross-equation restrictions.

Usage

SURest(data = data,
       eqlist = eqlist,
       restrictions = NULL,
       method="1wayWB")

Arguments

eqlist

a list containing the equations making up the SUR system. They should be object of the class "formula" and necessarly include the intercept,

restrictions

a vector containing constraints on the equation coefficients, which should be expressed in the form "equation_name$variable_name". Any spaces should be excluded from the restrictions definition. If one of the constraints includes an intercept term, the variable_name will be simply 'const'. Only simple restrictions involving equality between two parameters are considered, and not linear combinations involving more than two parameters,

method

the estimation method to be used, one of "1wayWB", "2wayWB", or "2wayQUE" (see details),

data

a data frame of the class "pdata.frame" (mandatory).

Details

SURest is a function for the GLS estimation of SUR system models on (unbalanced) panel data. It supports the following estimation methods: one-way error component procedure based on the Biorn (2004)’s procedure (1wayWB), two-way error component procedure based on the Biorn (2004)’s procedure (2wayWB), and the two-way QUE estimation procedure by Platoni et al. (2012) (2wayQUE).

Value

An object of class SURest, which is a list of the following elements:

Sigma_u

remainder error variance-covariance matrix,

Sigma_mu

individual error variance-covariance matrix,

Sigma_nu

time error variance-covariance matrix,

varnames

a vector whose elements are the names of the variables considered in the system equations, taking into account only the first appearance of those affected by restrictions on the coefficients,

Estimate

vector of the coefficient estimates of the system equations, taking into account only the first appearance of those affected by restrictions,

std_error

vector of the standard errors of the coefficient estimates, taking into account only the first appearance of those affected by restrictions,

tstat

vector of the t-statistics associated to the coefficient estimates, taking into account only the first appearance of those affected by restrictions,

pvalue

vector of the p-values associated to the t-statistics,

infoSample

information on the considered dataset obtained trougth the pdim command of the plm package,

neq

number of the system equations,

Rsquared

list of R-squared obtained for each equation of the estimated system,

method

method choosen for the system estimation.

References

Biorn E, (2004), Regression Systems for Unbalanced Panel Data: a Stepwise Maximum Likelihood Procedure, Journal of Econometrics, 122(2), 181–291.

Platoni S, Sckokai P, Moro D (2012), A Note on Two-way ECM Estimation of SUR Systems on Unbalanced Panel Data, Econometric Reviews, 31(2), 119–141.

Examples

data("SURdata", package="panelSUR")

## Data preparation
library(plm)
datap <- pdata.frame(data, index=c("IND", "TIME"))

## Equations specification
eq1<-Y1~X1+X2
eq2<-Y2~X1+X2+X3
eq3<-Y3~X2+X3
eqlist<-c(eq1,eq2,eq3)

## Constraints specification
constraints<-c("eq1$X2=eq2$X1","eq2$X3=eq3$X2")

## System estimation
mod1<-SURest(eqlist=eqlist,restrictions=constraints,method="2wayQUE",data=datap)