FEEMProbit: Estimate the individual fixed effects...

In walshc/FEEMProbit: FEEMProbit

Description

Function to estimate the individual fixed effects expecation-maximization estimator in Chen (2016)

Usage

FEEMProbit(formula, data, tol = 1e-9, show.progress = FALSE)

Arguments

formula	A symbolic description of the model to be estimated. The formula has three parts, the dependent variable, the independent variable and the fixed effect. For example, y ~ x1 + x2 | id, where id is the fixed effect
data	A data.frame.
tol	Tolerance level to use for the EM algorithm. Lower tolerances increase computation time.
show.progress	If TRUE, print the norm between parameter vectors in between each iteration.

Value

call	The function call.
coefficients	The estimated coefficients.
fixed.effects	The estimated individual fixed effects.
fitted.values	The fitted values of the model.
residuals	The model's residuals.
model	The data.frame used to estimate the model.

Examples

library(data.table)
set.seed(1)

# Create data:
N <- 200
nT <- 50
df <- data.table(id = rep(1:N, each = nT),
                 t  = rep(1:nT, N))

# Individual fixed effects:
alpha <- (1:N - N/2)/(N/2)
df[, alpha := rep(alpha, each = nT)]

# Explanatory variable:
df[, x := rnorm(N * nT, mean = -2, sd = 0.5)]

# Coefficient to be estimated:
beta <- 0.5

# Latent variable:
df[, ystar := beta * x + alpha + rnorm(N * nT)]

# Observed variable:
df[, y := as.numeric(ystar > 0)]

# Estimate the model:
FEEMProbit(y ~ x | id, data = df, tol = 1e-6)

walshc/FEEMProbit documentation built on May 3, 2019, 11:50 p.m.