FE_decompo: Decomposition of fixed-effects weights

In MatthieuStigler/multiDiff: Multi-year diff-diff and FE decompositions

Decomposition of fixed-effects weights

Description

Decomposition of fixed-effects weights

Usage

FE_decompo(
  data,
  y_var = "y",
  time.index = "Time",
  treat = "tr",
  unit.index = "unit",
  covar = NULL,
  fixed_effects = c("time", "unit", "both"),
  by = unit.index
)

Arguments

data	A data frame containing the variables of itnerest, should have names y_var, treat, time.index and unit.index
y_var	The name of the y variable
time.index	The name of the variable describing the time index (date, year, etc)
treat	The name of the treatment variable.
unit.index	The name of the units (individual, county, household, etc).
covar	covariate to include This will use the values indicated in time.index and unit.index
fixed_effects	The fixed effects to include, either time, unit ot both.
by	over which dimension to do it?

Examples

library(lfe)
 dat_sim_1 <- sim_dat(N=100, Time = 5)
coef_FE_time <- coef(felm(y~tr|Time, data = dat_sim_1))
coef_FE_unit <- coef(felm(y~tr|unit, data = dat_sim_1))

coefs_by_Y <- FE_decompo(data=dat_sim_1,
                         treat = "tr",
                         by = "Time",
                         fixed_effects = "time")
coefs_by_U <- FE_decompo(data=dat_sim_1,
                         treat = "tr",
                         by = "unit",
                         fixed_effects = "unit")
with(coefs_by_Y, weighted.mean(treat_coef, treat_weight))
coef_FE_time
with(coefs_by_U, weighted.mean(treat_coef, treat_weight))
coef_FE_unit

MatthieuStigler/multiDiff documentation built on Oct. 21, 2023, 8:13 a.m.