Table 3 in Kaboski and Townsend (2012), “The Impact of Credit on Village Economies”

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",
  "stargazer",
  "broom",
  "kableExtra",
  "scales",
  "magrittr"
)

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

# Load an example dataset ---------------
data_short <- read_dta("Kaboski2012_replication/AEJApp2009-0115_data/documentation/AnnualData_ShortSample.dta")

Data cleaning

table3_outcome_list <- c(
  "newst", "tc", "dlgassd", "dnetinc"
  )

data_table3 <- data_short %>% 
  group_by(case_id) %>% 
  mutate(
    farm = ifelse(
      !is.na(och),
      ifelse((och > -5) & (och <= 15), 1, 0),
      0
    ),
    age2h = ageh^2,
    vfst = vfst / 10000,
    villageyear = floor(case_id / 1000),
    invHH = ifelse(year == 1, lead(1 / vHH), 1 / vHH),
    invHHim = ifelse(year == 6, invHH, NA),
    invHHi = mean(invHHim, na.rm = TRUE),
    vHHi = 1 / invHHi,
    invHHpvf = invHHi * (year > 5),
    invHHtvf1 = invHHi * (year == 6),
    invHHtvf2 = invHHi * (year == 7),
    dlgassd = log(lead(gassd) / gassd),
    dnetinc = ifelse(
      is.infinite(log(lead(netinc) / netinc)),
      NA,
      log(lead(netinc) / netinc)
    )
  ) %>% 
  ungroup() %>% 
  filter(year < 8)

Create table

# list of functions to filter data to create subsample
table3_filter_list <- list(
  function (x) {filter(
    x, newst <= 315000,
    vHHi < 250, vHHi >= 50, year < 8
    )},
  function (x) {filter(
    x, tc >= 4430, tc <= 433000,
    vHHi < 250, vHHi >= 50, year < 8
    )},
  function (x) {filter(
    x, dlgassd >= -2.565758, dlgassd <= 2.494314,
    vHHi < 250, vHHi >= 50, year < 8
    )},
  function (x) {filter(
    x, dnetinc >= -2.605709, dnetinc <= 2.848096,
    vHHi < 250, vHHi >= 50, year < 8
    )}
  )

# function for regressions
reg_func <- function(varlist, filterlist, iv, data) {
  varlist %>% 
    enframe("model_no", "outcome") %>% 
    mutate(filter = filterlist) %>% 
    mutate(
      model = map2(outcome, filter, function(x, y) felm(
        formula(
          paste0(
            x, 
            "~ as.factor(year) + madult + fadult + kids + maleh + farm + ageh + age2h + educh | ",
            "case_id | ", iv, " | villageyear"
          )
        ),
        data = y(data)
      ))
    )
  }

# functions to arrange regression results for output table
table_arrange <- function(reg_output) {
  
  main_rhs = "`vfst(fit)`"
  
  reg_output %>% 
    mutate(
      coef = map_chr(model, function(x) ifelse(
        abs(x$coefficients[main_rhs, 1]) > 50,
        comma_format(accuracy = 1e+2)(x$coefficients[main_rhs, 1]),
        ifelse(
          abs(x$coefficients[main_rhs, 1]) > 5,
          comma_format(accuracy = 1e+1)(x$coefficients[main_rhs, 1]),
          comma_format(accuracy = 1e-3)(x$coefficients[main_rhs, 1])
        )
      )),
      pval = map_dbl(model, ~ .x$cpval[main_rhs]),
      coef = ifelse(
        pval < 0.01,
        paste0(coef, "***"),
        ifelse(
          pval < 0.05,
          paste0(coef, "**"),
          ifelse(
            pval < 0.10,
            paste0(coef, "*"),
            coef
          )
        )
      ),
      se = paste0(
        "(", 
        map_chr(model, function(x) ifelse(
          abs(x$cse[main_rhs]) > 50,
          comma_format(accuracy = 1e+2)(x$cse[main_rhs]),
          ifelse(
            abs(x$cse[main_rhs]) > 5,
            comma_format(accuracy = 1e+1)(x$cse[main_rhs]),
            comma_format(accuracy = 1e-3)(x$cse[main_rhs])
          )
        )),
        ")"
      )
    ) %>% 
    select(coef, se) %>% 
    as.matrix() %>% 
    t()
  }

# OLS regression
table3_ols_output <- reg_func(
  table3_outcome_list, 
  replicate(
    length(table3_outcome_list), 
    function(x) {filter(x, vHHi < 250, vHHi >= 50, year < 8)}
    ),
  "(vfst ~ vfst)",
  data_table3
  ) %>% 
  table_arrange()

# IV regression
table3_baseline_iv_output <- reg_func(
  table3_outcome_list, 
  replicate(
    length(table3_outcome_list), 
    function(x) {filter(x, vHHi < 250, vHHi >= 50, year < 8)}
    ),
  "(vfst ~ invHHtvf1 + invHHtvf2)",
  data_table3
  ) %>% 
  table_arrange()

# IV regression, all villages
table3_all_iv_output <- reg_func(
  table3_outcome_list, 
  replicate(
    length(table3_outcome_list), 
    function(x) {filter(x, year < 8)}
    ),
  "(vfst ~ invHHtvf1 + invHHtvf2)",
  data_table3
  ) %>% 
  table_arrange()

# IV regression with truncation
table3_no_outlier_iv_output <- reg_func(
  table3_outcome_list, 
  table3_filter_list,
  "(vfst ~ invHHtvf1 + invHHtvf2)",
  data_table3
  ) %>% 
  table_arrange()

# combine results for output table
c(
  "OLS regression", "",
  "Baseline IV regression: only villages", "\\hspace{1em}with 50-200 households",
  "IV regression using all villages", "",
  "IV regression without 1\\% outliers", ""
  ) %>% 
  cbind(
    rbind(
      table3_ols_output,
      table3_baseline_iv_output,
      table3_all_iv_output,
      table3_no_outlier_iv_output
    )
  ) %>% 
  as_tibble() %>% 
  mutate_all(linebreak, align = "l") %>% 
  set_names(NULL) %>% 
  kable(
    "latex", booktabs = TRUE, escape = FALSE, 
    align = c("l", rep("c", length(table3_outcome_list)))
    ) %>%
  kable_styling(latex_options = "scale_down") %>%
  add_header_above(
    c(
      "Response variable", "Net short-term\ncreidt level", 
      "Consumption\nlevel", "Asset\ngrowth rate", 
      "Net income\ngrowth rate"
      )
    ) %>%
  pack_rows(index = c(
    "Technique" = 8
    )) %>%
  save_kable("tex/Kaboski2012_table3_replicate.tex")