Native event-study estimator: relative-time dummies (with
reference_period dropped as the baseline) on unit and time
fixed effects, fitted by the same native TWFE engine as
morie_did_panel_fe. Reproduces
fixest::feols + fixest::i() to machine precision
(see tests/cross/).
Usage
morie_did_event_study(
data,
outcome,
unit,
time,
treatment_time,
covariates = NULL,
reference_period = -1L,
leads = 4L,
lags = 4L,
cluster = NULL,
alpha = 0.05
)Arguments
- data
Panel data frame.
- outcome
Outcome column.
- unit
Unit identifier column.
- time
Calendar-time column (integer-valued).
- treatment_time
Column giving the period in which each unit first received treatment (
InforNAfor never-treated units).- covariates
Optional time-varying covariates.
- reference_period
Relative-time period omitted as baseline (default
-1).- leads
Number of pre-treatment periods to include.
- lags
Number of post-treatment periods to include.
- cluster
Cluster variable for standard errors (defaults to
unit).- alpha
Significance level.
Value
A list with coefficients (data frame),
reference_period, pre_trend_f_stat,
pre_trend_p_value, and details.
Examples
set.seed(1)
df <- expand.grid(unit = 1:30, time = 1:6)
df$treat_time <- ifelse(df$unit <= 15, 4, Inf)
df$d <- as.integer(df$time >= df$treat_time)
df$y <- 0.1 * df$time + 0.7 * df$d + rnorm(nrow(df), sd = 0.4)
res <- morie_did_event_study(df, "y", "unit", "time", "treat_time",
leads = 2L, lags = 2L)
res$coefficients
#> relative_time estimate std_error ci_lower ci_upper p_value
#> 1 -2 -0.06545429 0.1456526 -0.35092815 0.2200196 6.531530e-01
#> 2 -1 0.00000000 0.0000000 0.00000000 0.0000000 NA
#> 3 0 0.50753098 0.1918843 0.13144473 0.8836172 8.169457e-03
#> 4 1 0.84043472 0.1951271 0.45799268 1.2228768 1.653977e-05
#> 5 2 0.40110683 0.1917268 0.02532915 0.7768845 3.643184e-02