Table 1B in Banerjee et al. (2015), “The Mircle of Microfinance? Evidence from a Randomized Evaluation”

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_endline <- read_dta("Banerjee2015_replication/2013-0533_data_endlines1and2.dta")

Table 1B

# outcome variables
hh_composition <- c(
  "hhsize", "adults", "children", "male_head", "head_age", "head_noeduc"
  )
credit_access <- c(
  "spandana", "othermfi", "anybank", "anyinformal", "anyloan"
  )
loan_amt <- c(
  "spandana_amt", "othermfi_amt", "bank_amt", "informal_amt", "anyloan_amt"
  )
self_emp_activ <- c(
  "total_biz", "female_biz", "female_biz_pct"
  )
businesses <- c(
  "bizrev", "bizexpense", "bizinvestment", 
  "bizemployees", "hours_week_biz"
  )
businesses_allHH <- c(
  "bizrev_allHH", "bizexpense_allHH", "bizinvestment_allHH", 
  "bizemployees_allHH", "hours_week_biz_allHH"
  )
consumption <- c(
  "total_exp_mo", "nondurable_exp_mo", "durables_exp_mo", "home_durable_index"
  )
all_varlist <- c(
  hh_composition, credit_access, loan_amt, self_emp_activ,
  businesses, businesses_allHH, consumption
  )

# data cleaning
data_table1B <- data_endline
for (i in seq(2)) {
  data_table1B <- data_table1B %>% 
    mutate(
      across(
        .cols = paste0(businesses, "_", as.character(i)),
        .fns = ~ .x,
        .names = paste0("{col}_allHH_", as.character(i))
      )
    ) %>% 
    mutate(
      across(
        .cols = paste0(businesses, "_", as.character(i)),
        .fns = ~ ifelse(eval(parse(text = paste0("total_biz_", as.character(i)))) == 0, NaN, .x)
      )
    ) %>% 
    rename_with(
      ~ str_replace(.x, paste0("_", as.character(i), "_allHH"), "_allHH")
    )
}

# variable labels
var_labels <- c(
  "Number members", "Number adults \n \\hspace{1em}($>=$ 16 years old)",
  "Number children \n \\hspace{1em}($<$ 16 years old)",
  "Male head", "Head's age", "Head with no education",
  "Loan from Spandana", "Loan from other MFI",
  "Loan from a bank", "Informal loan",
  "Any type of loan",
  "Spandana", "Other MFI", "Bank", "Informal loan",
  "Total",
  "Number of activities", 
  "Number of activities \n \\hspace{1em}managed by women",
  "Share of HH activities \n \\hspace{1em}managed by women",
  "Revenue / month (Rs)", "Expenses / month (Rs)",
  "Investment / month (Rs)", "Employment (employees)",
  "Self-employment (hours per week)",
  "Revenue / month (Rs)", "Expenses / month (Rs)",
  "Investment / month (Rs)", "Employment (employees)",
  "Self-employment (hours per week)",
  "Total consumption (Rs)", "Nondurables consumption (Rs)",
  "Durables consumption (Rs)", "Asset index"
)

# function to create a panel for control group
control_group_sumstat <- function(varlist, var_labels, data) {
  
  output <- data %>% 
    select(all_of(varlist)) %>% 
    map( ~ tibble(
      obs = format(sum(!is.na(.x)), digits = 0, format = "f", big.mark = ","),
      mean_var = mean(.x, na.rm = TRUE), 
      sd_var = sd(.x, na.rm = TRUE)
      )
      ) %>% 
    tibble() %>% 
    unnest(".") %>% 
    mutate(
      col_names = var_labels,
      mean_var = ifelse(
        mean_var >= 100, 
        formatC(mean_var, digits = 0, format = "f", big.mark = ","), 
        formatC(mean_var, digits = 3, format = "f", big.mark = ",")
        ),
      sd_var = paste0(
        "(", 
        ifelse(
          sd_var >= 100,
          formatC(sd_var, digits = 0, format = "f", big.mark = ","), 
          formatC(sd_var, digits = 3, format = "f", big.mark = ",")
        ),
        ")"
        )
      ) %>% 
    relocate(col_names)
  
  return(output)
}

# function to create a panel comparing two endlines
EL2_EL1_sumstat <- function(varlist, data) {
  
  data_long <- data %>% 
    select(c("treatment", "areaid", paste0(varlist, "_1"), paste0(varlist, "_2"))) %>% 
    pivot_longer(
      cols = c(paste0(varlist, "_1"), paste0(varlist, "_2")), 
      names_to = c(".value", "endline"),
      names_patter = "(.*)(_\\d$)"
      )
  
  output <- map(varlist, function(x) felm(
      formula(
          paste0(x, "~ as.factor(endline) | 0 | 0 | areaid"), 
        ),
      data = data_long %>% filter(treatment == 0))
    ) %>%
    enframe("model_no", "model") %>% 
    mutate(
      estimate = map_dbl(
        model, 
        function(x) summary(x)$coefficients["as.factor(endline)_2", "Estimate"]
        ),
      p_value = map_dbl(
        model, 
        function(x) waldtest(x, "as.factor(endline)_2")["p.F"]
        )
    ) %>% 
    select(estimate, p_value) %>% 
    mutate(
      across(
        .cols = c(estimate, p_value),
        .fns = function(x) ifelse(
          abs(x) > 10,
          formatC(x, digits = 0, format = "f", big.mark = ","),
          formatC(x, digits = 3, format = "f")
        )
      )
    )
  
  return(output)
}

# combine panels
cbind(
  control_group_sumstat(
    paste0(all_varlist, "_1"), 
    var_labels, 
    data_table1B %>% filter(treatment == 0)
    ) %>% 
    rename_with(.fn = function(x) paste0(x, "_1"), .cols = obs:sd_var),
  control_group_sumstat(
    paste0(all_varlist, "_2"), 
    var_labels, 
    data_table1B %>% filter(treatment == 0)
    ) %>% 
    rename_with(.fn = function(x) paste0(x, "_2"), .cols = obs:sd_var) %>% 
    select(-col_names),
  EL2_EL1_sumstat(all_varlist, data_table1B)
  ) %>% 
  mutate_all(linebreak, align = "l") %>% 
  set_colnames(NULL) %>%
  kable("latex", booktabs = TRUE, escape = FALSE, align = c("l", rep("r", 8))) %>%
  kable_styling(latex_options = "scale_down") %>%
  add_header_above(
    c(" ", rep(c("Obs.", "Mean", "SD"), 2), "Coeff.", "$p$-value"),
    escape = FALSE
    ) %>%
  add_header_above(
    c(
      " " = 1, "EL1 control group" = 3, 
      "EL2 control group" = 3, "EL2 - CL1" = 2
      )
    ) %>%
  pack_rows(index = c(
    "Household composition" = length(hh_composition),
    "Access to credit" = length(credit_access),
    "Amount borrowed from (in Rs)" = length(loan_amt),
    "Self-employment activities" = length(self_emp_activ),
    "Businesses" = length(businesses),
    "Businesses (all households)" = length(businesses_allHH),
    "Consumption (per household per month)" = length(consumption)
    )) %>%
  save_kable("tex/Banerjee_table1B_replicate.tex")