1.RUN_MODEL_VF - copia.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Librerias"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [],
   "source": [
    "#ARREGLAR GRÁFICAS, AÑADIR POBLACIÓN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [],
   "source": [
    "#INDEPENDIENTE "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 179,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [],
   "source": [
    "rm(list=ls())\n",
    "library(foreign)\n",
    "library(Synth)\n",
    "library(xtable)\n",
    "library(openxlsx)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 180,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [],
   "source": [
    "# POBLACIÓN\n",
    "# CAMBIAR VENTANA HACIA DELANTE (22)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 181,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [],
   "source": [
    "# Load Data \n",
    "df <- read.dta(\"dbec_vf.dta\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 182,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [],
   "source": [
    "df[\"POP\"] <- log(df[\"POP\"], base = 10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 183,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Warning message in dir.create(\"RESULTADOS_EXP\"):\n",
      "\"'RESULTADOS_EXP' already exists\"Warning message in dir.create(directorio):\n",
      "\"'RESULTADOS_EXP\\LACPER2010' already exists\""
     ]
    }
   ],
   "source": [
    "# year_t0 = 1995 #AÑO AL INICIO\n",
    "year_t0 = 2000 #AÑO AL INICIO\n",
    "year_t1 = 2010 #AÑO ANTES DEL TRATAMIENTO\n",
    "year_t2 = 2016 #AÑO FINAL\n",
    "\n",
    "d<-df[df$DATE<=year_t2,]\n",
    "d<-d[d$DATE>=year_t0,]\n",
    "\n",
    "\n",
    "pais = \"PER\"\n",
    "# pais = \"COL\"\n",
    "# pais = \"HND\"\n",
    "\n",
    "code_treated = d[d$ISO3==pais,]$WEO_CODE[1]\n",
    "\n",
    "\n",
    "muestra <- \"LAC\"\n",
    "# muestra <- \"CLU\"\n",
    "# muestra <- \"CLU&LAC\"\n",
    "\n",
    "if (muestra == \"LAC\") {\n",
    "    d <- d[d$LAC1 == 1, ]\n",
    "} else if (muestra == \"CLU\") {\n",
    "    d <- d[(d$CLUSTER>=4 | d$WEO==293) & (d$LAC1!=1 | d$WEO==293),]\n",
    "} else if (muestra == \"CLU&LAC\") {\n",
    "    d <- d[(d$CLUSTER>=3 | d$WEO==293) | (d$LAC1==1 | d$WEO==293),]\n",
    "}\n",
    "\n",
    "# Crear una carpeta\n",
    "dir.create(\"RESULTADOS_EXP\")\n",
    "directorio = paste0(\"RESULTADOS_EXP/\", muestra, pais, year_t1)\n",
    "dir.create(directorio)\n",
    "\n",
    "# d <- d[d$CLUSTER>=4 | d$WEO==293,]\n",
    "# d <- d[d$CLUSTER>=2 | d$WEO==293,]\n",
    "# d <- d[d$CLUSTER>=4 | d$LAC2==1,]\n",
    "# d <- d[d$GROUP==\"EMDEs\" | d$WEO==293,]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 184,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [],
   "source": [
    "d <- d[(d$WEO!=213),]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 185,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [],
   "source": [
    "d <- d[(d$WEO!=626),]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Run synthetic control"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 189,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "X1, X0, Z1, Z0 all come directly from dataprep object.\n",
      "\n",
      "\n",
      "**************** \n",
      " optimization over w weights: computing synthtic control unit \n",
      " \n",
      "\n",
      "\n",
      "**************** \n",
      "**************** \n",
      "**************** \n",
      "\n",
      "MSPE (LOSS V): 0.0001454332 \n",
      "\n",
      "solution.v:\n",
      " 1 \n",
      "\n",
      "solution.w:\n",
      " 0.006896748 0.01525705 0.007939943 0.003738041 0.001045109 0.005655355 0.009159505 0.04858983 0.005645635 0.005220568 0.8875503 0.003301911 \n",
      "\n"
     ]
    }
   ],
   "source": [
    "#################################################################################\n",
    "# PARAMETERS\n",
    "#################################################################################\n",
    "\n",
    "# predictors_dp       = c(\"PRIM_BAL_GDP\")\n",
    "# dependent_dp        = \"PRIM_BAL_GDP\"\n",
    "\n",
    "dep = \"INFLATION\"\n",
    "\n",
    "predictors_dp       = c(dep)\n",
    "dependent_dp        = dep\n",
    "\n",
    "\n",
    "id_variable         = which(names(d) == \"WEO_CODE\")\n",
    "time_variable       = which(names(d) == \"DATE\")\n",
    "\n",
    "# special_predictors  = list(list(\"TRADE_OPEN\", year_t0:year_t1, c(\"mean\")),\n",
    "#                             list(\"NET_FDI_GDP\", year_t0:year_t1, c(\"mean\")),\n",
    "#                             list(\"LUR\", year_t0:year_t1, c(\"mean\")),\n",
    "#                             list(\"CAP_RETURN\", year_t0:year_t1, c(\"mean\")),\n",
    "#                             list(\"POP\", year_t0:year_t1, c(\"mean\")),\n",
    "#                             list(\"INFLATION\", year_t0:year_t1, c(\"mean\"))\n",
    "#                            )\n",
    "\n",
    "special_predictors  = list(list(\"TRADE_OPEN\", year_t0:year_t1, c(\"mean\")),\n",
    "                            list(\"NET_FDI_GDP\", year_t0:year_t1, c(\"mean\")),\n",
    "                            list(\"LUR\", year_t0:year_t1, c(\"mean\")),\n",
    "                            list(\"CAP_RETURN\", year_t0:year_t1, c(\"mean\")),\n",
    "                            list(\"POP\", year_t0:year_t1, c(\"mean\")))\n",
    "\n",
    "treated             = code_treated \n",
    "controls            = setdiff(unique(d$WEO_CODE), treated) #EXLUYENDO AL TRATADO\n",
    "time_predictors     = year_t0:year_t1\n",
    "time_ssr            = year_t0:year_t1\n",
    "names               = which(names(d) == \"ISO3\")\n",
    "time_plot           = year_t0:year_t2\n",
    "\n",
    "#################################################################################\n",
    "# DATA PREPARATION\n",
    "#################################################################################\n",
    "\n",
    "\n",
    "dataprep.out <-\n",
    "  dataprep(\n",
    "           foo = d,\n",
    "           predictors    = predictors_dp,\n",
    "           dependent     = dependent_dp,\n",
    "           unit.variable = id_variable,\n",
    "           time.variable = time_variable,\n",
    "          #  special.predictors = special_predictors,\n",
    "           treatment.identifier = treated,\n",
    "           controls.identifier = controls,\n",
    "           time.predictors.prior = time_predictors,\n",
    "           time.optimize.ssr = time_ssr,\n",
    "           unit.names.variable = names,\n",
    "           time.plot = time_plot\n",
    "         )\n",
    "\n",
    "#################################################################################\n",
    "# FIT\n",
    "#################################################################################\n",
    "synth.out <- synth(\n",
    "        data.prep.obj=dataprep.out,\n",
    "        Margin.ipop=.005,Sigf.ipop=7,Bound.ipop=6\n",
    "        )\n",
    "\n",
    "synth.tables <- synth.tab(\n",
    "                          dataprep.res = dataprep.out,\n",
    "                          synth.res = synth.out\n",
    "                          )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 187,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [],
   "source": [
    "tabla_estimaciones <- createWorkbook()\n",
    "\n",
    "addWorksheet(tabla_estimaciones, \"variables\")\n",
    "pesos <- do.call(rbind.data.frame, synth.tables[2])\n",
    "descriptivas = do.call(rbind.data.frame, synth.tables[1])\n",
    "descriptivas <- cbind(row.names(descriptivas), descriptivas)\n",
    "descriptivas <- cbind(descriptivas, pesos)\n",
    "writeData(tabla_estimaciones, \"variables\", descriptivas)\n",
    "\n",
    "\n",
    "addWorksheet(tabla_estimaciones, \"countries\")\n",
    "paises = do.call(rbind.data.frame, synth.tables[3])\n",
    "writeData(tabla_estimaciones, \"countries\", paises)\n",
    "\n",
    "\n",
    "synthY0 <- (dataprep.out$Y0 %*% synth.out$solution.w)\n",
    "obsLAC  <- aggregate(d[\"DEBT_GDP\"], by=list(d$DATE), mean, na.rm=TRUE)[, 2]\n",
    "Y1      <- dataprep.out$Y1plot\n",
    "gap <- dataprep.out$Y1-(dataprep.out$Y0%*%synth.out$solution.w)\n",
    "\n",
    "fila <- rep(0, length(seq(year_t0, year_t2))); fila[seq(year_t0, year_t2) == year_t1] <- 1\n",
    "\n",
    "efecto <- data.frame(seq(year_t0, year_t2), ObsLAC = obsLAC, Y1 = Y1, SynthY0 = synthY0, Gap = gap, Fila = fila)\n",
    "colnames(efecto) <- c(\"Año\", \"Average Sample\", pais, paste0(pais, \"_synth\"), \"Effect\", \"Treatment Year\")\n",
    "addWorksheet(tabla_estimaciones, \"effects\")\n",
    "writeData(tabla_estimaciones, \"effects\", efecto)\n",
    "\n",
    "\n",
    "saveWorkbook(tabla_estimaciones, paste0(directorio, \"/Tablas.xlsx\"), overwrite = TRUE)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Resultados gráficos"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [],
   "source": [
    "max <- max(efecto[\"PER_synth\"])*1.1\n",
    "min <- min(efecto[\"PER_synth\"])*0.9\n",
    "\n",
    "# min <- 0\n",
    "\n",
    "\n",
    "# max <- 10\n",
    "# min <- 0\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [],
   "source": [
    "# min(efecto[\"PER_synth\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<strong>png:</strong> 2"
      ],
      "text/latex": [
       "\\textbf{png:} 2"
      ],
      "text/markdown": [
       "**png:** 2"
      ],
      "text/plain": [
       "png \n",
       "  2 "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<strong>png:</strong> 2"
      ],
      "text/latex": [
       "\\textbf{png:} 2"
      ],
      "text/markdown": [
       "**png:** 2"
      ],
      "text/plain": [
       "png \n",
       "  2 "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<strong>png:</strong> 2"
      ],
      "text/latex": [
       "\\textbf{png:} 2"
      ],
      "text/markdown": [
       "**png:** 2"
      ],
      "text/plain": [
       "png \n",
       "  2 "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#################################################################################\n",
    "# PERU VS SAMPLE\n",
    "#################################################################################\n",
    "\n",
    "\n",
    "options(repr.plot.width = 4, repr.plot.height = 4)\n",
    "cexl = 0.9\n",
    "par(cex = cexl, cex.axis = cexl, cex.lab = cexl, cex.main = cexl, cex.sub = cexl)\n",
    "\n",
    "png(paste0(directorio, paste0(\"/1_\", pais, \"_vs_sample.png\")), width = 600, height = 600)\n",
    "\n",
    "# Crear el gráfico\n",
    "plot(year_t0:year_t2, dataprep.out$Y1plot, type=\"l\", ylim=c(min, max), col=\"black\", lty=\"solid\",\n",
    "      ylab=\"Debt/GDP\",  xlab=\"Year\", xaxs=\"i\", yaxs=\"i\", lwd=2)\n",
    "\n",
    "lines(year_t0:year_t2, obsLAC , \n",
    "      col=\"black\", lty=\"dashed\", lwd=2)\n",
    "    \n",
    "abline(v=year_t1, lty=\"dotted\")\n",
    "\n",
    "legend(\"bottomleft\", legend = c(pais, \"sample\"),  lty = c(\"solid\", \"dashed\"), \n",
    "      col = c(\"black\", \"black\"), cex = 0.8, bg = \"white\")\n",
    "\n",
    "dev.off()\n",
    "\n",
    "#################################################################################\n",
    "# PERU VS SYTN PERÚ\n",
    "#################################################################################\n",
    "options(repr.plot.width = 4, repr.plot.height = 4)\n",
    "cexl = 0.9\n",
    "par(cex = cexl, cex.axis = cexl, cex.lab = cexl, cex.main = cexl, cex.sub = cexl)\n",
    "\n",
    "png(paste0(directorio, paste0(\"/2_\", pais, \"_vs_synth.png\")), width = 600, height = 600)\n",
    "\n",
    "# Crear el gráfico\n",
    "plot(year_t0:year_t2, Y1, type=\"l\", ylim=c(min, max), col=\"black\", lty=\"solid\",\n",
    "      ylab=\"Debt/GDP\",  xlab=\"Year\", xaxs=\"i\", yaxs=\"i\", lwd=2)\n",
    "\n",
    "lines(year_t0:year_t2,  synthY0, \n",
    "      col=\"black\", lty=\"dashed\", lwd=2)\n",
    "    \n",
    "abline(v=year_t1, lty=\"dotted\")\n",
    "\n",
    "legend(\"bottomleft\", legend = c(pais, \"_synth\"),  lty = c(\"solid\", \"dashed\"), \n",
    "      col = c(\"black\", \"black\"), cex = 0.8, bg = \"white\")\n",
    "\n",
    "dev.off()\n",
    "\n",
    "\n",
    "#################################################################################\n",
    "# EFECTO\n",
    "#################################################################################\n",
    "options(repr.plot.width = 4, repr.plot.height = 4)\n",
    "cexl = 0.9\n",
    "par(cex = cexl, cex.axis = cexl, cex.lab = cexl, cex.main = cexl, cex.sub = cexl)\n",
    "\n",
    "png(paste0(directorio, paste0(\"/3_\", pais, \"_effect.png\")), width = 600, height = 600)\n",
    "\n",
    "plot(year_t0:year_t2,gap,\n",
    "type=\"l\",ylim=c(-0.3,0.3),col=\"black\",lty=\"solid\",\n",
    "ylab =\"DEBT_DGP\",\n",
    "     xlab =\"Year\",\n",
    "     xaxs = \"i\", yaxs = \"i\",\n",
    "     lwd=2)\n",
    "     \n",
    "abline(h=0,lty=\"dotted\")\n",
    "abline(v=2010,lty=\"dotted\")\n",
    "\n",
    "dev.off()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Inference"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Leave-one-out to construct p-values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "X1, X0, Z1, Z0 all come directly from dataprep object.\n",
      "\n",
      "\n",
      "**************** \n",
      " searching for synthetic control unit  \n",
      " \n",
      "\n",
      "**************** \n",
      "**************** \n",
      "**************** \n",
      "\n",
      "MSPE (LOSS V): 0.0375686 \n",
      "\n",
      "solution.v:\n",
      " 0.1671745 0.1696513 0.001938978 0.000122222 1.0689e-06 0.3135096 0.3106206 0.01311903 0.02380727 5.54684e-05 \n",
      "\n",
      "solution.w:\n",
      " 7.7417e-06 2.7e-09 3e-10 1.12e-08 3e-10 1.5e-08 1.41e-08 6.2473e-06 5.4e-09 8.9e-09 0.6665388 0.3332442 3.9e-09 2.05e-08 1.72e-08 0.000202711 2.032e-07 3e-10 2e-10 \n",
      "\n",
      "\n",
      "X1, X0, Z1, Z0 all come directly from dataprep object.\n",
      "\n",
      "\n",
      "**************** \n",
      " searching for synthetic control unit  \n",
      " \n",
      "\n",
      "**************** \n",
      "**************** \n",
      "**************** \n",
      "\n",
      "MSPE (LOSS V): 0.003745111 \n",
      "\n",
      "solution.v:\n",
      " 0.1980136 0.2039251 0.00932821 0.002493241 0.1282822 0.09639159 0.1388686 0.01166542 0.01455146 0.1964806 \n",
      "\n",
      "solution.w:\n",
      " 4.642e-07 3e-10 0.3325846 5e-10 3.5e-09 1.8e-09 2.9e-09 1.8e-09 6e-10 7e-10 2e-09 3e-10 2e-10 0.6674149 4e-10 6e-10 0 2e-10 1e-10 \n",
      "\n",
      "\n",
      "X1, X0, Z1, Z0 all come directly from dataprep object.\n",
      "\n",
      "\n",
      "**************** \n",
      " searching for synthetic control unit  \n",
      " \n",
      "\n",
      "**************** \n",
      "**************** \n",
      "**************** \n",
      "\n",
      "MSPE (LOSS V): 0.02577876 \n",
      "\n",
      "solution.v:\n",
      " 0.4798011 0.001291424 0.01792141 0.2098768 0.04034682 0.07233378 0.05906844 0.0168303 0.09598882 0.006541089 \n",
      "\n",
      "solution.w:\n",
      " 0 3e-10 2.2e-09 1.8e-09 3e-10 2e-10 1.8e-09 3e-10 2e-10 4e-10 5e-10 0 1 1e-10 0 0 0 2e-10 1e-10 \n",
      "\n",
      "\n",
      "X1, X0, Z1, Z0 all come directly from dataprep object.\n",
      "\n",
      "\n",
      "**************** \n",
      " searching for synthetic control unit  \n",
      " \n",
      "\n",
      "**************** \n",
      "**************** \n",
      "**************** \n",
      "\n",
      "MSPE (LOSS V): 9.480143e-05 \n",
      "\n",
      "solution.v:\n",
      " 0.5044209 0.294286 0.1686411 1.188e-07 0.01147892 0.0002370925 0.001954642 0.003274837 0.01321111 0.002495298 \n",
      "\n",
      "solution.w:\n",
      " 3.539e-07 0.4118379 0.388945 0.1552609 4.84e-08 5.85e-08 2.89e-08 1.272e-07 5.176e-07 8.8026e-06 1.561e-07 1.103e-07 0.001945048 1.018e-07 1.815e-07 8.87199e-05 1.658e-07 0.04191133 4.434e-07 \n",
      "\n",
      "\n",
      "X1, X0, Z1, Z0 all come directly from dataprep object.\n",
      "\n",
      "\n",
      "**************** \n",
      " searching for synthetic control unit  \n",
      " \n",
      "\n",
      "**************** \n",
      "**************** \n",
      "**************** \n",
      "\n",
      "MSPE (LOSS V): 0.0001717439 \n",
      "\n",
      "solution.v:\n",
      " 0.1258426 0.07667351 0.02216472 0.1117316 0.1506366 0.02256723 0.09741616 0.1541894 0.08730925 0.1514691 \n",
      "\n",
      "solution.w:\n",
      " 1e-09 2e-10 1e-10 2e-10 0.09366881 0.3288141 1.9e-09 3.2e-09 0.02234162 2e-10 0.05684556 8e-10 0.4983299 2e-10 4e-10 2e-10 3e-10 4e-10 2e-10 \n",
      "\n",
      "\n",
      "X1, X0, Z1, Z0 all come directly from dataprep object.\n",
      "\n",
      "\n",
      "**************** \n",
      " searching for synthetic control unit  \n",
      " \n",
      "\n",
      "**************** \n",
      "**************** \n",
      "**************** \n",
      "\n",
      "MSPE (LOSS V): 0.006523001 \n",
      "\n",
      "solution.v:\n",
      " 0.3008662 0.5291273 0.07768473 5.9212e-06 0.03705194 0.001483691 0.001088753 0.05134012 0.0009482542 0.0004031064 \n",
      "\n",
      "solution.w:\n",
      " 1.03e-08 1.86e-08 0.211394 0.120341 7.6e-09 0.6257861 1.1e-08 3.6e-09 5.36e-08 8.7e-09 9.6e-09 4.9e-09 5.4e-09 9.04726e-05 1.23e-08 0.04231905 6.5e-09 9.98e-07 6.82296e-05 \n",
      "\n",
      "\n",
      "X1, X0, Z1, Z0 all come directly from dataprep object.\n",
      "\n",
      "\n",
      "**************** \n",
      " searching for synthetic control unit  \n",
      " \n",
      "\n",
      "**************** \n",
      "**************** \n",
      "**************** \n",
      "\n",
      "MSPE (LOSS V): 0.009109453 \n",
      "\n",
      "solution.v:\n",
      " 0.4371637 0.2673022 0.1745215 0.0476754 0.07052476 0.0004674484 0.0001352388 0.001678932 1.32145e-05 0.000517562 \n",
      "\n",
      "solution.w:\n",
      " 1.98e-08 0.02270686 1.09e-08 0.169342 0.127989 0.5553536 1.324e-07 2.77e-08 1.35e-08 2.45e-08 2.74e-08 0.1246079 2.94e-08 1.69e-08 6.74e-08 2.61e-08 1.97e-08 1.488e-07 5.71e-08 \n",
      "\n",
      "\n",
      "X1, X0, Z1, Z0 all come directly from dataprep object.\n",
      "\n",
      "\n",
      "**************** \n",
      " searching for synthetic control unit  \n",
      " \n",
      "\n",
      "**************** \n",
      "**************** \n",
      "**************** \n",
      "\n",
      "MSPE (LOSS V): 0.0005620597 \n",
      "\n",
      "solution.v:\n",
      " 0.1195032 0.1139732 0.1012721 6.75059e-05 0.3353837 0.0002540328 0.0001997862 0.02907607 0.2507216 0.04954885 \n",
      "\n",
      "solution.w:\n",
      " 3.35e-08 0.2481084 6e-09 0.1792411 0.131327 2.037e-07 1.5948e-06 3.1e-09 7.268e-07 1.1e-08 6.3e-09 2.3e-09 0.3392382 4.56e-08 9.5e-09 7.9e-09 7.8e-09 0.1020825 2.3e-08 \n",
      "\n",
      "\n",
      "X1, X0, Z1, Z0 all come directly from dataprep object.\n",
      "\n",
      "\n",
      "**************** \n",
      " searching for synthetic control unit  \n",
      " \n",
      "\n",
      "**************** \n",
      "**************** \n",
      "**************** \n",
      "\n",
      "MSPE (LOSS V): 0.001267441 \n",
      "\n",
      "solution.v:\n",
      " 0.488711 0.2609735 0.1491168 0.03614486 0.001043702 0.0001852826 0.009292927 0.004402519 0.04970654 0.0004228662 \n",
      "\n",
      "solution.w:\n",
      " 1.209e-07 0.08424996 2.16e-08 1.65e-08 1.58e-08 7.6e-09 7.7e-09 1.06e-08 6e-09 0.636043 2.32e-07 0.03390342 0.06978862 0.09569514 1.56e-08 1.12e-08 0.08031934 1.18e-08 1.94e-08 \n",
      "\n",
      "\n",
      "X1, X0, Z1, Z0 all come directly from dataprep object.\n",
      "\n",
      "\n",
      "**************** \n",
      " searching for synthetic control unit  \n",
      " \n",
      "\n",
      "**************** \n",
      "**************** \n",
      "**************** \n",
      "\n",
      "MSPE (LOSS V): 0.007031661 \n",
      "\n",
      "solution.v:\n",
      " 0.1445824 0.0006130293 0.1286509 0.01338727 0.168872 0.0002963048 0.1168129 0.05715458 0.2101847 0.1594459 \n",
      "\n",
      "solution.w:\n",
      " 4.9e-09 2e-09 1.6e-09 2.1e-09 0.101208 0.2883668 0.1697776 2.4e-08 2.98e-08 2.6698e-06 5.9e-09 3.65e-07 0.4406445 2.7e-09 9.7e-09 9.4e-09 6.5e-09 2.5e-09 2.3e-09 \n",
      "\n",
      "\n",
      "X1, X0, Z1, Z0 all come directly from dataprep object.\n",
      "\n",
      "\n",
      "**************** \n",
      " searching for synthetic control unit  \n",
      " \n",
      "\n",
      "**************** \n",
      "**************** \n",
      "**************** \n",
      "\n",
      "MSPE (LOSS V): 0.0006007577 \n",
      "\n",
      "solution.v:\n",
      " 0.3380754 0.1734689 0.1210178 0.2112657 2.02535e-05 0.000168609 0.03697011 0.07976721 0.03548633 0.003759667 \n",
      "\n",
      "solution.w:\n",
      " 6.37e-08 1.116e-07 0.1386992 6.024e-07 2.68e-08 2.17e-08 2.412e-07 5.638e-07 0.4539167 0.3231313 6.74e-08 0.01993759 0.06431062 1.73e-07 6.337e-07 1.146e-07 2.332e-07 1.122e-06 6.785e-07 \n",
      "\n",
      "\n",
      "X1, X0, Z1, Z0 all come directly from dataprep object.\n",
      "\n",
      "\n",
      "**************** \n",
      " searching for synthetic control unit  \n",
      " \n",
      "\n",
      "**************** \n",
      "**************** \n",
      "**************** \n",
      "\n",
      "MSPE (LOSS V): 0.003199444 \n",
      "\n",
      "solution.v:\n",
      " 0.1699805 0.0009668291 0.0008295073 0.01189487 0.001971304 0.02774554 0.2402671 0.1220361 0.256266 0.1680423 \n",
      "\n",
      "solution.w:\n",
      " 0.4842472 0.06149871 8.248e-07 0.03213896 0.2808604 3.494e-06 4.87e-08 7.15e-08 0.008062939 8.01e-08 9.34e-08 7.9939e-06 1.123e-07 9.94e-08 8.99e-08 1.3015e-06 4.37e-08 0.1331774 3.19e-08 \n",
      "\n",
      "\n",
      "X1, X0, Z1, Z0 all come directly from dataprep object.\n",
      "\n",
      "\n",
      "**************** \n",
      " searching for synthetic control unit  \n",
      " \n",
      "\n",
      "**************** \n",
      "**************** \n",
      "**************** \n",
      "\n",
      "MSPE (LOSS V): 0.02548877 \n",
      "\n",
      "solution.v:\n",
      " 0.3329107 1.3052e-06 6.5803e-06 0.1656022 3.898e-07 0.01832058 0.05031582 0.2581301 0.06156286 0.1131495 \n",
      "\n",
      "solution.w:\n",
      " 6.20279e-05 5.71518e-05 0 2.08e-08 8e-10 2.9e-09 1.1e-09 6e-10 0.0176822 2.4e-09 2.5e-09 6.52e-08 2e-10 0.635007 1.187e-07 2.873e-07 0.347185 2.16e-06 4.0264e-06 \n",
      "\n",
      "\n",
      "X1, X0, Z1, Z0 all come directly from dataprep object.\n",
      "\n",
      "\n",
      "**************** \n",
      " searching for synthetic control unit  \n",
      " \n",
      "\n",
      "**************** \n",
      "**************** \n",
      "**************** \n",
      "\n",
      "MSPE (LOSS V): 0.00275671 \n",
      "\n",
      "solution.v:\n",
      " 0.3675635 0.06220389 0.002454552 0.1437641 0.007596532 0.09151948 0.02377098 2.94725e-05 0.2846567 0.0164408 \n",
      "\n",
      "solution.w:\n",
      " 0 0 0.6045689 0 2.9871e-06 0 1e-10 1e-10 0 5e-09 5.854e-07 0 0 0 0 0 3e-10 0.3954275 1.18e-08 \n",
      "\n",
      "\n",
      "X1, X0, Z1, Z0 all come directly from dataprep object.\n",
      "\n",
      "\n",
      "**************** \n",
      " searching for synthetic control unit  \n",
      " \n",
      "\n",
      "**************** \n",
      "**************** \n",
      "**************** \n",
      "\n",
      "MSPE (LOSS V): 0.006007952 \n",
      "\n",
      "solution.v:\n",
      " 0.422643 0.2674301 0.2006667 4.1447e-06 0.09688397 3.7798e-06 9.93279e-05 0.0004962464 0.002074942 0.009697799 \n",
      "\n",
      "solution.w:\n",
      " 0.1328972 0.6198883 3.33e-08 9.65e-08 5.69e-08 7.52e-08 1.48e-07 9.83e-08 3.1053e-06 8.36e-08 8.96e-08 6.89923e-05 0.1531963 5.35e-08 5.662e-07 6.61e-08 0.09394401 1.37e-07 5.62e-07 \n",
      "\n",
      "\n",
      "X1, X0, Z1, Z0 all come directly from dataprep object.\n",
      "\n",
      "\n",
      "**************** \n",
      " searching for synthetic control unit  \n",
      " \n",
      "\n",
      "**************** \n",
      "**************** \n",
      "**************** \n",
      "\n",
      "MSPE (LOSS V): 0.02403525 \n",
      "\n",
      "solution.v:\n",
      " 0.1131484 0.1004835 0.1455165 0.003488925 0.1220031 1.20192e-05 9.78891e-05 0.1343144 0.1449048 0.2360305 \n",
      "\n",
      "solution.w:\n",
      " 2.1e-09 3.4e-09 6e-10 1.9e-09 2.32e-08 3.227e-07 4.8157e-06 2.7e-09 3e-10 5.8e-09 5e-10 4.3e-09 0.1265265 6e-10 0 2.59e-08 0.2665769 8.76e-08 0.6068913 \n",
      "\n",
      "\n",
      "X1, X0, Z1, Z0 all come directly from dataprep object.\n",
      "\n",
      "\n",
      "**************** \n",
      " searching for synthetic control unit  \n",
      " \n",
      "\n",
      "**************** \n",
      "**************** \n",
      "**************** \n",
      "\n",
      "MSPE (LOSS V): 0.01586506 \n",
      "\n",
      "solution.v:\n",
      " 0.2400648 0.09670395 0.1059539 0.02631247 1.75841e-05 0.2655191 0.2255284 0.03944263 0.0001639318 0.0002931758 \n",
      "\n",
      "solution.w:\n",
      " 2.1e-09 0.04898119 1.8e-09 1.12e-08 1.2e-09 0.105323 1e-10 5e-10 8e-10 1.2e-09 1.2e-09 2.3e-09 0.1645478 1.3e-09 0.2366235 2.343e-07 0.4362108 9.52e-08 0.008313325 \n",
      "\n",
      "\n",
      "X1, X0, Z1, Z0 all come directly from dataprep object.\n",
      "\n",
      "\n",
      "**************** \n",
      " searching for synthetic control unit  \n",
      " \n",
      "\n",
      "**************** \n",
      "**************** \n",
      "**************** \n",
      "\n",
      "MSPE (LOSS V): 0.1353037 \n",
      "\n",
      "solution.v:\n",
      " 0.4830233 0.2858152 0.1663415 0.009423159 0.01831941 0.01952211 0.004744801 0.003917066 0.002500827 0.006392634 \n",
      "\n",
      "solution.w:\n",
      " 9e-10 2e-10 3e-10 4e-10 5e-10 4e-10 5e-10 5e-10 7e-10 4e-10 5e-10 7e-10 0.09451222 4e-10 4e-10 1.4e-09 0.9054878 7e-10 7e-10 \n",
      "\n",
      "\n",
      "X1, X0, Z1, Z0 all come directly from dataprep object.\n",
      "\n",
      "\n",
      "**************** \n",
      " searching for synthetic control unit  \n",
      " \n",
      "\n",
      "**************** \n",
      "**************** \n",
      "**************** \n",
      "\n",
      "MSPE (LOSS V): 0.003896032 \n",
      "\n",
      "solution.v:\n",
      " 0.1715145 0.02253307 0.343014 0.0007229985 0.002521562 3.077e-07 0.2381797 2.13128e-05 0.1053817 0.1161109 \n",
      "\n",
      "solution.w:\n",
      " 3.94e-08 8.2e-09 1.27e-08 2.05e-08 4.647e-07 3.07e-08 1.89e-08 1.68e-08 2.34e-08 7.44e-08 2.14e-08 3.81e-08 0.01672477 0.2628899 7.9e-09 2.85e-08 3.85e-08 2.78e-08 0.7203845 \n",
      "\n",
      "\n",
      "X1, X0, Z1, Z0 all come directly from dataprep object.\n",
      "\n",
      "\n",
      "**************** \n",
      " searching for synthetic control unit  \n",
      " \n",
      "\n",
      "**************** \n",
      "**************** \n",
      "**************** \n",
      "\n",
      "MSPE (LOSS V): 0.00321026 \n",
      "\n",
      "solution.v:\n",
      " 0.3713837 0.2173592 7.47414e-05 0.01187598 0.2697301 0.0001317277 0.04590835 0.01066009 0.04883304 0.02404311 \n",
      "\n",
      "solution.w:\n",
      " 6.217e-07 0.1426659 9.8e-08 2.982e-07 1.149e-07 1.581e-07 1.291e-07 1.293e-07 1.728e-07 1.118e-07 1.687e-07 2.966e-07 0.07540196 9.23e-08 1.6863e-06 4.881e-07 8.204e-07 2.9857e-06 0.7819237 \n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Asumiendo que d es tu dataframe y ya está cargado\n",
    "\n",
    "# Definir los años\n",
    "year_t0 = 2000 # Año al inicio\n",
    "year_t1 = 2010 # Año antes del tratamiento\n",
    "year_t2 = 2015 # Año final\n",
    "\n",
    "# Lista de tratados\n",
    "lista_tratados = unique(d$WEO_CODE)\n",
    "\n",
    "# Preparar un dataframe para guardar los resultados\n",
    "resultados = data.frame()\n",
    "\n",
    "resultados_lista = list()\n",
    "\n",
    "# Bucle para cada país o código en la lista de tratados\n",
    "for (i in lista_tratados) {\n",
    "    # Definir variables para dataprep\n",
    "    # predictors_dp = c(\"DEBT_GDP\",\"EXP_GDP\")\n",
    "    # dependent_dp = \"DEBT_GDP\"\n",
    "    # id_variable = which(names(d) == \"WEO_CODE\")\n",
    "    # time_variable = which(names(d) == \"DATE\")\n",
    "    # special_predictors = list(list(\"REV_GDP\", year_t0:year_t1, c(\"mean\")))\n",
    "    # treated = i\n",
    "    # controls = setdiff(unique(d$WEO_CODE), treated)\n",
    "    # time_predictors = year_t0:year_t1\n",
    "    # time_ssr = year_t0:year_t1\n",
    "    # names_variable = which(names(d) == \"ISO3\")\n",
    "    # time_plot = year_t0:year_t2\n",
    "\n",
    "\n",
    "    predictors_dp       = c(\"DEBT_GDP\",\"EXP_GDP\")\n",
    "    dependent_dp        = \"DEBT_GDP\"\n",
    "    id_variable         = which(names(d) == \"WEO_CODE\")\n",
    "    time_variable       = which(names(d) == \"DATE\")\n",
    "\n",
    "    special_predictors  = list(list(\"REV_GDP\", year_t0:year_t1, c(\"mean\")),\n",
    "                            #    list(\"WGI_CONTR_CORR\", year_t0:year_t1, c(\"mean\")),\n",
    "                            #    list(\"WGI_GOV_EFFECT\", year_t0:year_t1, c(\"mean\")),\n",
    "                            #    list(\"WGI_POL_STAB\", year_t0:year_t1, c(\"mean\")),\n",
    "                            list(\"WGI_REG_QUAL\", year_t0:year_t1, c(\"mean\")),\n",
    "                            list(\"TRADE_OPEN\", year_t0:year_t1, c(\"mean\")),\n",
    "                            #     list(\"CONSUMPTION_PC\", year_t0:year_t1, c(\"mean\")),\n",
    "                                list(\"NET_FDI_GDP\", year_t0:year_t1, c(\"mean\")),\n",
    "                            #  list(\"BROAD_MONEY_GDP\", year_t0:year_t1, c(\"mean\")),\n",
    "                                list(\"LUR\", year_t0:year_t1, c(\"mean\")),\n",
    "                                list(\"CAP_RETURN\", year_t0:year_t1, c(\"mean\")),\n",
    "                            #  list(\"INT_RESERVES_CAP\", year_t0:year_t1, c(\"mean\")),\n",
    "                                list(\"INFLATION\", year_t0:year_t1, c(\"mean\")),\n",
    "                                # list(\"BANK_CRISIS\", year_t0:year_t1, c(\"mean\")),\n",
    "                                list(\"COMDEP\", year_t0:year_t1, c(\"mean\"))) #OTROS PREDICTORES (EN LA MEDIA)\n",
    "\n",
    "\n",
    "    # special_predictors  = list(list(\"REV_GDP\", year_t0:year_t1, c(\"mean\"))) #OTROS PREDICTORES (EN LA MEDIA)\n",
    "    treated             = i\n",
    "    controls            = setdiff(unique(d$WEO_CODE), treated) #EXLUYENDO AL TRATADO\n",
    "    time_predictors     = year_t0:year_t1\n",
    "    time_ssr            = year_t0:year_t1\n",
    "    names               = which(names(d) == \"ISO3\")\n",
    "    time_plot           = year_t0:year_t2\n",
    "\n",
    "\n",
    "\n",
    "    # Ejecutar dataprep\n",
    "    dataprep.out <- dataprep(\n",
    "        foo = d,\n",
    "        predictors = predictors_dp,\n",
    "        dependent = dependent_dp,\n",
    "        unit.variable = id_variable,\n",
    "        time.variable = time_variable,\n",
    "        special.predictors = special_predictors,\n",
    "        treatment.identifier = treated,\n",
    "        controls.identifier = controls,\n",
    "        time.predictors.prior = time_predictors,\n",
    "        time.optimize.ssr = time_ssr,\n",
    "        unit.names.variable = names_variable,\n",
    "        time.plot = time_plot\n",
    "    )\n",
    "\n",
    "    # Ejecutar synth\n",
    "    synth.out <- synth(\n",
    "        data.prep.obj = dataprep.out,\n",
    "        Margin.ipop = .005,\n",
    "        Sigf.ipop = 7,\n",
    "        Bound.ipop = 6\n",
    "    )\n",
    "\n",
    "    gap <- dataprep.out$Y1-(dataprep.out$Y0%*%synth.out$solution.w)\n",
    "\n",
    "    # Guardar los resultados\n",
    "    # Crear un dataframe temporal\n",
    "\n",
    "    nombre = as.character(i)\n",
    "\n",
    "    temp_df <- data.frame(nombre = gap)\n",
    "    \n",
    "    resultados_lista[[nombre]] <- temp_df\n",
    "\n",
    "}\n",
    "\n",
    "# Ver los resultados\n",
    "resultados <- do.call(cbind, resultados_lista)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [],
   "source": [
    "pvalues = list() \n",
    "controles = setdiff(unique(names(resultados)), \"X293\")\n",
    "\n",
    "for (i in controles) {\n",
    "\n",
    "    r = as.numeric(resultados[[\"X293\"]]<resultados[[i]])\n",
    "    nombre = as.character(i)\n",
    "    temp_df <- data.frame(nombre = r)  \n",
    "    pvalues[[nombre]] <- temp_df\n",
    "\n",
    "}\n",
    "\n",
    "pvalues <- do.call(cbind, pvalues)\n",
    "colnames(pvalues) <- controles\n",
    "p_values_emp = 1- colSums(t(pvalues))/ncol(pvalues)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<ol class=list-inline>\n",
       "\t<li>0.736842105263158</li>\n",
       "\t<li>0.526315789473684</li>\n",
       "\t<li>0.210526315789474</li>\n",
       "\t<li>0.736842105263158</li>\n",
       "\t<li>0.736842105263158</li>\n",
       "\t<li>0.421052631578947</li>\n",
       "\t<li>0.421052631578947</li>\n",
       "\t<li>0.368421052631579</li>\n",
       "\t<li>0.526315789473684</li>\n",
       "\t<li>0.421052631578947</li>\n",
       "\t<li>0.421052631578947</li>\n",
       "\t<li>0.315789473684211</li>\n",
       "\t<li>0.315789473684211</li>\n",
       "\t<li>0.263157894736842</li>\n",
       "\t<li>0.210526315789474</li>\n",
       "\t<li>0.315789473684211</li>\n",
       "</ol>\n"
      ],
      "text/latex": [
       "\\begin{enumerate*}\n",
       "\\item 0.736842105263158\n",
       "\\item 0.526315789473684\n",
       "\\item 0.210526315789474\n",
       "\\item 0.736842105263158\n",
       "\\item 0.736842105263158\n",
       "\\item 0.421052631578947\n",
       "\\item 0.421052631578947\n",
       "\\item 0.368421052631579\n",
       "\\item 0.526315789473684\n",
       "\\item 0.421052631578947\n",
       "\\item 0.421052631578947\n",
       "\\item 0.315789473684211\n",
       "\\item 0.315789473684211\n",
       "\\item 0.263157894736842\n",
       "\\item 0.210526315789474\n",
       "\\item 0.315789473684211\n",
       "\\end{enumerate*}\n"
      ],
      "text/markdown": [
       "1. 0.736842105263158\n",
       "2. 0.526315789473684\n",
       "3. 0.210526315789474\n",
       "4. 0.736842105263158\n",
       "5. 0.736842105263158\n",
       "6. 0.421052631578947\n",
       "7. 0.421052631578947\n",
       "8. 0.368421052631579\n",
       "9. 0.526315789473684\n",
       "10. 0.421052631578947\n",
       "11. 0.421052631578947\n",
       "12. 0.315789473684211\n",
       "13. 0.315789473684211\n",
       "14. 0.263157894736842\n",
       "15. 0.210526315789474\n",
       "16. 0.315789473684211\n",
       "\n",
       "\n"
      ],
      "text/plain": [
       " [1] 0.7368421 0.5263158 0.2105263 0.7368421 0.7368421 0.4210526 0.4210526\n",
       " [8] 0.3684211 0.5263158 0.4210526 0.4210526 0.3157895 0.3157895 0.2631579\n",
       "[15] 0.2105263 0.3157895"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "p_values_emp"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Ratio RPMSE Post and Pre treatment"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [],
   "source": [
    "rmse <- function(x){sqrt(mean(x^2))}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<dl class=dl-horizontal>\n",
       "\t<dt>X213</dt>\n",
       "\t\t<dd>0.193826218269806</dd>\n",
       "\t<dt>X223</dt>\n",
       "\t\t<dd>0.0611973142288493</dd>\n",
       "\t<dt>X228</dt>\n",
       "\t\t<dd>0.160557646367397</dd>\n",
       "\t<dt>X233</dt>\n",
       "\t\t<dd>0.00973660247345846</dd>\n",
       "\t<dt>X238</dt>\n",
       "\t\t<dd>0.0131051110704865</dd>\n",
       "\t<dt>X248</dt>\n",
       "\t\t<dd>0.0807651001733466</dd>\n",
       "\t<dt>X268</dt>\n",
       "\t\t<dd>0.0954434558138238</dd>\n",
       "\t<dt>X273</dt>\n",
       "\t\t<dd>0.0237077976604233</dd>\n",
       "\t<dt>X283</dt>\n",
       "\t\t<dd>0.0356011343894374</dd>\n",
       "\t<dt>X288</dt>\n",
       "\t\t<dd>0.0838550013100424</dd>\n",
       "\t<dt>X293</dt>\n",
       "\t\t<dd>0.0245103597022144</dd>\n",
       "\t<dt>X298</dt>\n",
       "\t\t<dd>0.0565636289065356</dd>\n",
       "\t<dt>X311</dt>\n",
       "\t\t<dd>0.159652036419961</dd>\n",
       "\t<dt>X313</dt>\n",
       "\t\t<dd>0.0525043808291194</dd>\n",
       "\t<dt>X321</dt>\n",
       "\t\t<dd>0.0775109814210906</dd>\n",
       "\t<dt>X328</dt>\n",
       "\t\t<dd>0.155033059603155</dd>\n",
       "\t<dt>X343</dt>\n",
       "\t\t<dd>0.125956596041443</dd>\n",
       "\t<dt>X361</dt>\n",
       "\t\t<dd>0.367836447426575</dd>\n",
       "\t<dt>X362</dt>\n",
       "\t\t<dd>0.0624182022542134</dd>\n",
       "\t<dt>X364</dt>\n",
       "\t\t<dd>0.0566591605268501</dd>\n",
       "</dl>\n"
      ],
      "text/latex": [
       "\\begin{description*}\n",
       "\\item[X213] 0.193826218269806\n",
       "\\item[X223] 0.0611973142288493\n",
       "\\item[X228] 0.160557646367397\n",
       "\\item[X233] 0.00973660247345846\n",
       "\\item[X238] 0.0131051110704865\n",
       "\\item[X248] 0.0807651001733466\n",
       "\\item[X268] 0.0954434558138238\n",
       "\\item[X273] 0.0237077976604233\n",
       "\\item[X283] 0.0356011343894374\n",
       "\\item[X288] 0.0838550013100424\n",
       "\\item[X293] 0.0245103597022144\n",
       "\\item[X298] 0.0565636289065356\n",
       "\\item[X311] 0.159652036419961\n",
       "\\item[X313] 0.0525043808291194\n",
       "\\item[X321] 0.0775109814210906\n",
       "\\item[X328] 0.155033059603155\n",
       "\\item[X343] 0.125956596041443\n",
       "\\item[X361] 0.367836447426575\n",
       "\\item[X362] 0.0624182022542134\n",
       "\\item[X364] 0.0566591605268501\n",
       "\\end{description*}\n"
      ],
      "text/markdown": [
       "X213\n",
       ":   0.193826218269806X223\n",
       ":   0.0611973142288493X228\n",
       ":   0.160557646367397X233\n",
       ":   0.00973660247345846X238\n",
       ":   0.0131051110704865X248\n",
       ":   0.0807651001733466X268\n",
       ":   0.0954434558138238X273\n",
       ":   0.0237077976604233X283\n",
       ":   0.0356011343894374X288\n",
       ":   0.0838550013100424X293\n",
       ":   0.0245103597022144X298\n",
       ":   0.0565636289065356X311\n",
       ":   0.159652036419961X313\n",
       ":   0.0525043808291194X321\n",
       ":   0.0775109814210906X328\n",
       ":   0.155033059603155X343\n",
       ":   0.125956596041443X361\n",
       ":   0.367836447426575X362\n",
       ":   0.0624182022542134X364\n",
       ":   0.0566591605268501\n",
       "\n"
      ],
      "text/plain": [
       "       X213        X223        X228        X233        X238        X248 \n",
       "0.193826218 0.061197314 0.160557646 0.009736602 0.013105111 0.080765100 \n",
       "       X268        X273        X283        X288        X293        X298 \n",
       "0.095443456 0.023707798 0.035601134 0.083855001 0.024510360 0.056563629 \n",
       "       X311        X313        X321        X328        X343        X361 \n",
       "0.159652036 0.052504381 0.077510981 0.155033060 0.125956596 0.367836447 \n",
       "       X362        X364 \n",
       "0.062418202 0.056659161 "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "preloss"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "plot without title"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "anios <- rownames(resultados)\n",
    "t0 <- which(anios == year_t0)\n",
    "t1 <- which(anios == year_t1)\n",
    "t2 <- which(anios == year_t2)\n",
    "\n",
    "\n",
    "\n",
    "preloss  <- apply(resultados[t0:t1,],2,rmse)\n",
    "postloss <- apply(resultados[t0:t2,],2,rmse)\n",
    "\n",
    "dotchart(sort(postloss/preloss),\n",
    "         xlab=\"Post-Period RMSE / Pre-Period RMSE\",\n",
    "         pch=19)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "vscode": {
     "languageId": "r"
    }
   },
   "outputs": [],
   "source": [
    "special_predictors  = list(list(\"REV_GDP\", year_t0:year_t1, c(\"mean\")),\n",
    "                        #    list(\"WGI_CONTR_CORR\", year_t0:year_t1, c(\"mean\")),\n",
    "                        #    list(\"WGI_GOV_EFFECT\", year_t0:year_t1, c(\"mean\")),\n",
    "                        #    list(\"WGI_POL_STAB\", year_t0:year_t1, c(\"mean\")),\n",
    "                           # list(\"WGI_REG_QUAL\", year_t0:year_t1, c(\"mean\")),\n",
    "                           list(\"TRADE_OPEN\", year_t0:year_t1, c(\"mean\")),\n",
    "                        #     list(\"CONSUMPTION_PC\", year_t0:year_t1, c(\"mean\")),\n",
    "                            list(\"NET_FDI_GDP\", year_t0:year_t1, c(\"mean\")),\n",
    "                           #  list(\"BROAD_MONEY_GDP\", year_t0:year_t1, c(\"mean\")),\n",
    "                            list(\"LUR\", year_t0:year_t1, c(\"mean\")),\n",
    "                            list(\"CAP_RETURN\", year_t0:year_t1, c(\"mean\")),\n",
    "                           #  list(\"INT_RESERVES_CAP\", year_t0:year_t1, c(\"mean\")),\n",
    "                            list(\"INFLATION\", year_t0:year_t1, c(\"mean\"))\n",
    "                           #  list(\"POB\", year_t0:year_t1, c(\"mean\"))\n",
    "                           #  list(\"POP\", year_t0:year_t1, c(\"mean\"))\n",
    "                           #  list(\"COMDEP\", year_t0:year_t1, c(\"mea"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "R",
   "language": "R",
   "name": "ir"
  },
  "language_info": {
   "codemirror_mode": "r",
   "file_extension": ".r",
   "mimetype": "text/x-r-source",
   "name": "R",
   "pygments_lexer": "r",
   "version": "3.6.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}