Library: | panel |
See also: | panfix panhaus pandyn |
Quantlet: | panhaus | |
Description: | Hausman test for correlation between the regressors and the individual effects |
Usage: | output=panhaus(Z,siga,sige,m {,T} ) | |
Input: | ||
Z | NT x (k+2) matrix, data set ordered by individuals (i), 1st column: index of individuals 2nd column: dependent variable 3rd, ..., k-th column: explanatory variables (without constant!) | |
siga | scalar, variance estimate of individual effects | |
sige Definiton scalar, variance estimate of the remaining error | scalar, number of explanatory variables which are correlated with the fixed effects, the remaining variables are assumed to be uncorrelated with the individual effects (= mixed specification) | |
m | optional scalar, number of time periods for a BALANCED data set. There has to be no column for the index of individuals. (for a data set with the same T for all individuals = BALANCED PANEL) | |
Output: | ||
output | string with output table |
library("panel") z=read("earnings") z1 = z[,1:2]~log(z[,3])~z[,4]~(z[,4]^2) {output,siga,sige}=panfix(z1,2) panhaus(z1,siga,sige,2)
Contents of output [ 1,] "=====================================================" [ 2,] "Hausman Specification Test for the null hypothesis: " [ 3,] " 2 Variables x(i,t) are uncorrelated with a(i) " [ 4,] "=====================================================" [ 5,] " d = beta(between)-beta(within) SE t-value" [ 6,] "=====================================================" [ 7,] " d [ 1 ] = -0.0705 0.0402 -1.754" [ 8,] " d [ 2 ] = 0.0001 0.0008 0.113" [ 9,] "=====================================================" [10,] "P-Value (RANDOM against FIXED effects): 0.0000" [11,] "====================================================="