Chamberlain's conditional (fixed-effects) logit, whose likelihood conditions on
each unit's number of ones (denominator by dynamic programming), and
random-effects logit or probit with a normal unit effect integrated by
Gauss-Hermite quadrature; maximised by BFGS with standard errors from the
numerical Hessian.
Usage
PanelBinaryChoice(
y,
X,
group,
model = c("fe_logit", "re_logit", "re_probit"),
n_quad = 30
)
Arguments
- y
Binary outcomes.
- X
Covariate matrix (no intercept column).
- group
Unit labels.
- model
"fe_logit", "re_logit" or "re_probit".
- n_quad
Quadrature nodes.
Value
list(coef, se, sigma, loglik, n_units, converged).
References
Chamberlain, G. (1980). Analysis of covariance with qualitative
data. Review of Economic Studies 47, 225-238. Butler, J. S. and Moffitt, R.
(1982). A computationally efficient quadrature procedure for the one-factor
multinomial probit model. Econometrica 50, 761-764.
Examples
PanelBinaryChoice(c(0, 1, 1, 0, 0, 1), matrix(c(0, 1, 2, 0.5, 0.2, 1.5)), c(1, 1, 1, 2, 2,
2))$n_units
#> [1] 2
