fixest (version 0.5.1)

dof: Type of degree of freedom in fixest summary

Description

Provides how the degrees of freedom should be calculated in vcov.fixest/summary.fixest.

Usage

dof(adj = 1, fixef.K = "nested", fixef.exact = FALSE, cluster.adj = TRUE)

Arguments

adj

Can be equal to 0, 1 or 2. Type of small sample correction. If 0: no small sample correction. If 1: a correction of the form n / (n - K) with n the number of observation and K the number of estiamted parameters. If 2: the correction is (n - 1) / (n - K).

fixef.K

How to account for the fixed-effects parameters, defaults to "nested". If FALSE or "no", fixed-effects parameters are discarded, meaning the number of parameters is only equal to the number of variables. If TRUE or yes, then the number of parameters is equal to the number of variables plus the number of fixed-effects. Finally, if nested, then the number of parameters is equal to the number of variables an the number of fixed-effects that *are not* nested in the clusters used to cluster the standard-errors.

fixef.exact

Logical, default is FALSE. If there are 2 or more fixed-effects, these fixed-effects they can be irregular, meaning they can provide the same information. If so, the "real" number of parameters should be lower than the total number of fixed-effects. If fixef.exact = TRUE, then fixef.fixest is first run to determine the exact number of parameters among the fixed-effects. Mostly, panels of the type individual-firm-year require fixef.exact = TRUE (but it adds computational costs).

cluster.adj

Logical, default is TRUE. How to make the small sample correction when clustering the standard-errors? If TRUE a G/(G-1) correction is performed with G the number of cluster values.

Value

It returns a dof.type object.

See Also

summary.fixest, vcov.fixest, setFixest_dof

Examples

# NOT RUN {
#
# Equivalence with lm/glm standard-errors
#

res = feols(Petal.Length ~ Petal.Width + Species, iris)
res_lm = lm(Petal.Length ~ Petal.Width + Species, iris)

# lm applies a correction of the type adj = 1 (fixest's default)
vcov(res) / vcov(res_lm)

# Glm
res_pois = fepois(round(Petal.Length) ~ Petal.Width + Species, iris)
res_glm = glm(round(Petal.Length) ~ Petal.Width + Species, iris, family = poisson())

# glm applies a correction of the type adj = 0 (this time not fixest's default)
vcov(res_pois, dof = dof(adj = 0)) / vcov(res_glm)

# Same example with the Gamma
res_gamma = feglm(round(Petal.Length) ~ Petal.Width + Species, iris, family = Gamma())
res_glm_gamma = glm(round(Petal.Length) ~ Petal.Width + Species, iris, family = Gamma())

vcov(res_gamma, dof = dof(adj = 0)) / vcov(res_glm_gamma)

#
# Fixed-effects corrections
#

# We create "irregular" FEs
base = data.frame(x = rnorm(10))
base$y = base$x + rnorm(10)
base$fe1 = rep(1:3, c(4, 3, 3))
base$fe2 = rep(1:5, each = 2)

est = feols(y ~ x | fe1 + fe2, base)

# fe1: 3 FEs
# fe2: 5 FEs

#
# Clustered standard-errors: by fe1
#

# Default: fixef.K = "nested"
#  => adjustment K = 1 + 5 (i.e. x + fe2)
summary(est)
attributes(vcov(est))[c("dof.type", "dof.K")]


# fixef.K = FALSE
#  => adjustment K = 1 (i.e. only x)
summary(est, dof = dof(fixef.K=FALSE))
attr(vcov(est, dof = dof(fixef.K=FALSE)), "dof.K")


# fixef.K = TRUE
#  => adjustment K = 1 + 3 + 5 - 1 (i.e. x + fe1 + fe2 - 1 restriction)
summary(est, dof = dof(fixef.K=TRUE))
attr(vcov(est, dof = dof(fixef.K=TRUE)), "dof.K")


# fixef.K = TRUE & fixef.exact = TRUE
#  => adjustment K = 1 + 3 + 5 - 2 (i.e. x + fe1 + fe2 - 2 restrictions)
summary(est, dof = dof(fixef.K=TRUE, fixef.exact = TRUE))
attr(vcov(est, dof = dof(fixef.K=TRUE, fixef.exact = TRUE)), "dof.K")

# There are two restrictions:
attr(fixef(est), "references")


# }