Table 2 in Fafchamps et al. (2014), “Microenterprise growth and the flypaper effect: Evidence from a randomized experiment in Ghana”

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

• float
• pdflscape
• booktabs
• multicol

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

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

# Load an example dataset ---------------
data <- read_dta("data/ReplicationDataGhanaJDE.dta")

# Add variables
data <- data %>%
  mutate(realfinalprofit_wave1 = ifelse(wave == 1, realfinalprofit, NA)) %>%
  group_by(sheno) %>%
  mutate(
    realfinalprofit_wave1 = max(realfinalprofit_wave1, na.rm = TRUE),
    realfinalprofit_wave1 = ifelse(is.infinite(realfinalprofit_wave1), NA, realfinalprofit_wave1)
    ) %>%
  ungroup()

# variable list
var_list <- c(
  "realfinalprofit", "fem", "highcapture", "highcapital", 
  "male_male", "male_mixed", "female_female", "female_mixed", 
  "realfinalprofit_wave1", "finalsales", "hourslastweek", 
  "totalK", "inventories", "useasusu", 
  "businesshome", "firmage", "everloan", "business_taxnumber", 
  "married", "educ_years", "digitspan", "akanspeaker", 
  "gaspeaker", "age"
)

# variable labels
var_label <- c(
  "Monthly profits in January 2009", "Female", 
  "High capture", "High baseline capital stock", 
  "Male in male dominated industry", "Male in mixed industry", 
  "Female in female dominated industry", "Female in mixed industry",
  "Monthly profits in October/November 2008", "Monthly sales in January 2009", 
  "Number of hours worked in last week", "Total capital stock in January 2009", 
  "Inventories at end of January 2009", "Uses a susu collector",
  "Business operated out of home", "Age of firm", 
  "Ever had bank or microfinance loan", "Business has a tax number",
  "Owner is married", "Owner's years of education", 
  "Owner's digitspan Recall", "Owner is Akan speaker",
  "Owner is Ga/Dangme speaker", "Owner's age"
)

# function to create a table for subsample
table2_sub <- function(data) {
  
  # regress outcome on control, cash treatment, and in-kind treatment
  # to get their means
  df_model <- map(var_list, function(x) lm(
    as.formula(
      paste(x, paste(
        c("-1", "control", "cashtreat", "equiptreat"), collapse = " + "
        ), sep = " ~ ")
      ), data = filter(data, wave == 2))
    ) %>%
    enframe("model_no", "model")
  
  # extract coefficients from the regression results
  # and keep as matrix form
  df_est <- df_model %>%
    mutate(tidied = map(model, tidy)) %>%
    select(model_no, tidied) %>%
    unnest(cols = c(tidied)) %>%
    mutate_if(is.double, formatC, digits = 2, format = "f") %>%
    select(model_no, term, estimate) %>%
    spread(term, estimate) %>%
    select(-model_no) %>% 
    as.matrix()
    
  # combine the table with sample size and p-value of F-test
  cbind(
    map_int(df_model$model, function(x) nrow(x$model)),
    df_est,
    map_chr(
      df_model$model, function(x) formatC(
        waldtest(x, ~ control - cashtreat | control - equiptreat)["p.F"],
        digits = 3, format = "f"
        )
      )
  )
}

# combine the variable labels and two summary statistics tables for 
# two subsamples
cbind(
  var_label, 
  table2_sub(data), 
  table2_sub(filter(data, is.na(trimgroup)))
  ) %>%
  set_colnames(NULL) %>%
  kable("latex", booktabs = TRUE) %>%
  kable_styling(latex_options = "scale_down") %>%
  add_header_above(
    c(" ", rep(c("N", "Mean", "Mean", "Mean", "$p$-value"), 2)),
    escape = FALSE
    ) %>%
  add_header_above(
    c(" ", rep(c(" ", "Control", "Cash", "In-kind", " "), 2))
    ) %>%
  add_header_above(c(" " = 1, "Full sample" = 5, "Trimmed sample" = 5)) %>%
  pack_rows(index = c(
    "Variables using to stratify or match" = 8,
    "Other variables" = length(var_list) - 8
    )) %>%
  save_kable("tex/FMQW2014_table2_replicate.tex")