refactored preprocessing of welfare effects data, added comments

scripts/paramspaces.R

-# This R code generates random draws from a +-10% interval around the 
-# elasticities of substitution esubd(i) for each sector in the Pothen and
-# Huebler model (2018). It then writes these draws to a CSV file to be 
-# used as the parameter space for sensitivity analysis in Snakemake.
+# This R code generates parameter spaces for the distributional
+# sensivitity analyses of the main results of Hübler et al. (2022).
+# More specifically, 1000 random draws from a +-10% interval around 
+# each of the sector-specific elasiticity estimates are generated.
+# This results in 1000 sets of sectoral parameter values to be used
+# in a parameter sweep in Snakemake.
+#
+# A separate parameter space is generated for each of the three sets
+# of sector-specific elasticity parameters present in the model:
+# - esubd(i): elasticities between domestically produced versus
+#   imported goods
+# - esubm(i): Armington elasticities
+# - esubva(j): elasticities between production factors
 # 
-# Marius Braun, May 2022
+# Marius Braun, April 2023
 
+# load packages
+library(ggplot2)
 library(readr)
+library(extrafont)
+library(openxlsx)
+library(Rcpp)
+library(tictoc)
+library(moments)
+library(confintr)
+library(stringr)
 library(dplyr)
 
+# clear workspace
 rm(list = ls())
 
 num_draws = 1000  # number of draws from interval
 set.seed(127)   # set random seed
-variation = 0.1
-num_decimals = 3
+variation = 0.1 # size of variation
+num_decimals = 3 # max. number of decimals (important due to file path length restrictions)
 
 # create vector of elasticity names
 elasticities = c("esubd", "esubm", "esubva")
+
 for(i in 1:length(elasticities)) {
   # load CSV file with baseline parameter values
   filename = paste("intervals_10p", elasticities[i], sep = "_")
@@ -58,5 +78,4 @@ for(i in 1:length(elasticities)) {
     row.names = FALSE
     )
 }
-rm(i, j, filename, file, paramspace)
-  
+rm(i, j, filename, file, paramspace)
\ No newline at end of file
\ No newline at end of file

scripts/welf_data.R

+# This R code extracts welfare effects from the output files of the
+# distributional sensitivity analysis of the main results of Hübler
+# et al. (2022). Welfare effects are then stored in a separate data
+# frame for esubd(i), esubm(i) and esubva(j) and saved in RDS files
+# to be used in subsequent analyses.
+#
+# Marius Braun, April 2023
+
+# load packages
+library(ggplot2)
+library(readr)
+library(extrafont)
+library(openxlsx)
+library(Rcpp)
+library(tictoc)
+library(moments)
+library(confintr)
+library(stringr)
+library(dplyr)
+
+# clear workspace
+rm(list = ls())
+
+# this is where you put the output files from the sensitivity analysis
+# (use separate subdirectories for esubd(i), esubm(i), and esubva(j))
+dir = "C:/Users/Marius Braun/output_sensitivity"
+
+# string vectors for elasticity, policy and income group names
+elasticities = c("esubd", "esubm", "esubva")
+policies = c("policy", "cbam")
+inc_groups = c("lo", "mi", "hi")
+
+for(i in 1:length(elasticities)) {
+  # load file paths of output files
+  filenames = as.data.frame(
+    list.files(
+      paste(
+        dir,
+        elasticities[i],
+        sep = "/"
+        ),
+      pattern = "output.xlsx",
+      full.names = T,
+      recursive = T)
+  )
+  colnames(filenames) = "files"
+  
+  # create empty data frame to store welfare effects
+  welf_name = paste(
+    "welf",
+    elasticities[i],
+    sep = "_"
+  )
+  welf = as.data.frame(
+    matrix(,
+           nrow = nrow(filenames),
+           ncol = length(policies) * length(inc_groups)
+    )
+  )
+  
+  # name columns of data frame according to policy and income group
+  for(j in 1:length(policies)) {
+    for(k in 1:length(inc_groups)) {
+      col_num = k +
+        (as.numeric(policies[j] == "cbam") * length(inc_groups))
+      colnames(welf)[col_num] = paste(
+        policies[j],
+        inc_groups[k],
+        sep = "_"
+      )
+    }
+  }
+  
+  # get welfare effects from output files
+  for(l in 1:nrow(filenames)) {
+    f = filenames$files[l]
+    if(file.exists(f)) {
+      welfare = read.xlsx(f, sheet = "welfp") %>%
+        filter(TOC == "DEU" &
+                 str_detect(X3,
+                            paste(
+                              paste0("\\b",
+                                     policies),
+                              collapse = "|")
+                 )
+        )
+      welfare = welfare %>% arrange(desc(X3))
+    }
+    welf[l, ] = t(welfare$X4)
+  }
+  
+  # name welfare effects data frame
+  assign(
+    x = welf_name,
+    value = welf
+  )
+  
+  # save welfare effects data frame as RDS file
+  saveRDS(get(welf_name),
+          paste0(
+            "prepared/",
+            welf_name,
+            ".rds")
+          )
+}
+rm(i, j, k, l, f, col_num, welf_name, welf, filenames,
+   welfare, dir, elasticities, inc_groups, policies)
\ No newline at end of file

setup.R

@@ -20,7 +20,7 @@ renv::init(bare = TRUE)
 # Install the packages
 install.packages(c(
   "ggplot2", "readr", "extrafont", "openxlsx", "Rcpp", "tictoc", "moments",
-  "confintr", "dplyr"
+  "confintr", "dplyr", "stringr"
   ))
 
 # Take a snapshot of the renv