Table 2 in Bandiera et al. (2020), “Women’s Empowerment in Action: Evidence from a Randomized Control Trial in Africa”

To read the created .tex file, the following \(\LaTeX\) packages are required:

• float
• pdflscape
• booktabs
• multicol
• makecell

# Install and load packages ---------------
packages <- c(
  "tidyverse",
  "haven",
  "lfe",
  "broom",
  "stargazer",
  "kableExtra",
  "magrittr"
)

pacman::p_load(packages, character.only = TRUE, install = TRUE)

# Load an example dataset ---------------
data <- read_dta("Bandiera2020_replication/ELA.dta")
data <- data %>% 
  mutate(
    follow_endline = ifelse(
      follow_up == 1 & endline == 1, 1, 0
      )
    )

# RHS variables
var_list_lin <- c(
   "age", "E_Denrolled", "partner","M_children", 
   "igaALL", "control_bodyALL", "aspirationALL"
   )
var_list <- c(var_list_lin, paste0(var_list_lin, ":treatment"))

# variable labels
var_label <- c(
  "Treatment",
  "Age", 
  "\\makecell[l]{Currently enrolled in\\\\\\hspace{1em}school}",
  "Married or cohabiting", "Has child(ren)", 
  "\\makecell[l]{Econonic\\\\\\hspace{1em}empowerment index}",
  "\\makecell[l]{Control over the\\\\\\hspace{1em}body index}",
  "Aspiration index",
  "Treatment $\\times$ age", 
  "\\makecell[l]{Treatment $\\times$ currently\\\\\\hspace{1em}enrolled inschool}",
  "\\makecell[l]{Treatment $\\times$ married\\\\\\hspace{1em}or cohabiting}", 
  "\\makecell[l]{Treatment $\\times$ has\\\\\\hspace{1em}child(ren)}", 
  "\\makecell[l]{Treatment $\\times$ econonic\\\\\\hspace{1em}empowerment index}",
  "\\makecell[l]{Treatment $\\times$ control\\\\\\hspace{1em}over the body index}",
  "\\makecell[l]{Treatment $\\times$ aspiration\\\\\\hspace{1em}index}"
)

# outcome variables
outcome_list <- c(
  rep("follow_up", 3),
  rep("endline", 3),
  rep("follow_endline", 3)
)

# list of control variables
control_list <- vector(mode = "list", length = length(outcome_list))
for (i in c(3, 6, 9)) {
  control_list[[i]] <- var_list
}

# fixed effects used in each regression
fe_list <- rep(c("0", "branch_name", "branch_name"), 3)

# regression results
reg_res <- pmap(
  list(outcome_list, control_list, fe_list),
  function(outcome, control, fe) felm(
    as.formula(
      paste0(
        outcome, 
        "~ ", paste0(c("treatment", control), collapse = " + "),
        "| ", fe, " | 0 | villid"
      )
    ),
    data = data
  )
)
# mean of outcome variables
mean_dep <- pmap_chr(
  list(outcome_list, reg_res, seq(9)),
  function(x, y, z) ifelse(
    z %in% c(2, 5, 8),
    formatC(colMeans(model.frame(y)[x], na.rm = TRUE), digits = 3, format = "f"),
    ""
    )
  )
branch_dummies <- rep(c("No", "Yes", "Yes"), 3)

# F-test that all interactions have 0 coefficients
fstat_list <- c()
for (i in seq(length(outcome_list))) {
  if (i %in% c(3, 6, 9)) {
  test_res <- waldtest(reg_res[[i]], paste0("treatment:", var_list_lin))[c("F", "p.F")] %>% 
    formatC(digits = 3, format = "f") %>% 
    paste0(collapse = "/")
    fstat_list <- c(fstat_list, test_res)
  } else {
    fstat_list <- c(fstat_list, " ")
  }
}

# table output
reg_res %>% 
  stargazer(
    dep.var.labels.include = FALSE,
    covariate.labels = var_label,
    keep = c("treatment", var_list),
    add.lines = list(
      c("\\makecell[l]{Mean of dependent\\\\\\hspace{1em}variable}", mean_dep),
      c("Branch dummies", branch_dummies),
      c("$F$-statistic [$p$-value]", fstat_list)
    ),
    title = "",
    type = "latex",
    out = "tex/Bandiera2020_table2_replicate.tex",
    omit.stat = c("adj.rsq", "rsq", "ser"),
    table.layout = "=#c-t-sa-n",
    digits = 3
  )