apoorvalal/MHE_py.ipynb

## MHE_py.ipynb
{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "#%%\n",
    "# pyscience imports\n",
    "import os\n",
    "import sys\n",
    "import glob\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import statsmodels.api as sm\n",
    "import statsmodels.formula.api as smf\n",
    "\n",
    "# graphics\n",
    "from plotnine import *\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "# plt.style.use(\"seaborn-darkgrid\")\n",
    "sns.set(style=\"ticks\", context=\"talk\")\n",
    "plt.style.use(\"dark_background\") # dark bg plots\n",
    "\n",
    "# %matplotlib inline\n",
    "# run for jupyter notebook\n",
    "from IPython.core.interactiveshell import InteractiveShell\n",
    "InteractiveShell.ast_node_interactivity = 'all'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'/home/alal/Desktop/code/mostly-harmless-replication'"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "%pwd"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Regressions "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import urllib\n",
    "import zipfile\n",
    "import urllib.request"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "dldat = False\n",
    "if dldat:\n",
    "    # Download data and unzip \n",
    "    urllib.request.urlretrieve('http://economics.mit.edu/files/397', 'asciiqob.zip')\n",
    "    with zipfile.ZipFile('asciiqob.zip', \"r\") as z:\n",
    "       z.extractall()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>lwklywge</th>\n",
       "      <th>educ</th>\n",
       "      <th>yob</th>\n",
       "      <th>qob</th>\n",
       "      <th>pob</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5.790019</td>\n",
       "      <td>12</td>\n",
       "      <td>30</td>\n",
       "      <td>1</td>\n",
       "      <td>45</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5.952494</td>\n",
       "      <td>11</td>\n",
       "      <td>30</td>\n",
       "      <td>1</td>\n",
       "      <td>45</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>5.315949</td>\n",
       "      <td>12</td>\n",
       "      <td>30</td>\n",
       "      <td>1</td>\n",
       "      <td>45</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>5.595926</td>\n",
       "      <td>12</td>\n",
       "      <td>30</td>\n",
       "      <td>1</td>\n",
       "      <td>45</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>6.068915</td>\n",
       "      <td>12</td>\n",
       "      <td>30</td>\n",
       "      <td>1</td>\n",
       "      <td>37</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   lwklywge  educ  yob  qob  pob\n",
       "0  5.790019    12   30    1   45\n",
       "1  5.952494    11   30    1   45\n",
       "2  5.315949    12   30    1   45\n",
       "3  5.595926    12   30    1   45\n",
       "4  6.068915    12   30    1   37"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Read the data into a pandas dataframe\n",
    "pums         = pd.read_csv('asciiqob.txt',\n",
    "                           header           = None,\n",
    "                           delim_whitespace = True)\n",
    "pums.columns = ['lwklywge', 'educ', 'yob', 'qob', 'pob']\n",
    "pums.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:               lwklywge   R-squared:                       0.117\n",
      "Model:                            OLS   Adj. R-squared:                  0.117\n",
      "Method:                 Least Squares   F-statistic:                 4.378e+04\n",
      "Date:                Fri, 04 Jan 2019   Prob (F-statistic):               0.00\n",
      "Time:                        22:53:12   Log-Likelihood:            -3.1935e+05\n",
      "No. Observations:              329509   AIC:                         6.387e+05\n",
      "Df Residuals:                  329507   BIC:                         6.387e+05\n",
      "Df Model:                           1                                         \n",
      "Covariance Type:            nonrobust                                         \n",
      "==============================================================================\n",
      "                 coef    std err          t      P>|t|      [0.025      0.975]\n",
      "------------------------------------------------------------------------------\n",
      "Intercept      4.9952      0.004   1118.882      0.000       4.986       5.004\n",
      "educ           0.0709      0.000    209.243      0.000       0.070       0.072\n",
      "==============================================================================\n",
      "Omnibus:                   191064.440   Durbin-Watson:                   1.870\n",
      "Prob(Omnibus):                  0.000   Jarque-Bera (JB):          4082110.366\n",
      "Skew:                          -2.377   Prob(JB):                         0.00\n",
      "Kurtosis:                      19.575   Cond. No.                         53.3\n",
      "==============================================================================\n",
      "\n",
      "Warnings:\n",
      "[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.\n"
     ]
    }
   ],
   "source": [
    "model      = smf.ols('lwklywge ~ educ', data=pums)\n",
    "print(model.fit().summary())\n",
    "#%%\n",
    "results    = model.fit()\n",
    "educ_coef  = results.params[1]\n",
    "intercept  = results.params[0]\n",
    "#%%"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Calculate means by educ attainment and predicted values\n",
    "groupbyeduc = pums.groupby('educ')\n",
    "educ_means  = groupbyeduc['lwklywge'].mean()\n",
    "yhat        = pd.Series(intercept + educ_coef * educ_means.index.values,\n",
    "                        index = educ_means.index.values)\n",
    "#%%"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>educ</th>\n",
       "      <th>lwklywge</th>\n",
       "      <th>intercept</th>\n",
       "      <th>yhat</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>5.022035</td>\n",
       "      <td>1</td>\n",
       "      <td>4.995182</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>5.064200</td>\n",
       "      <td>1</td>\n",
       "      <td>5.066033</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>5.166692</td>\n",
       "      <td>1</td>\n",
       "      <td>5.136884</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>5.173006</td>\n",
       "      <td>1</td>\n",
       "      <td>5.207735</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>5.264427</td>\n",
       "      <td>1</td>\n",
       "      <td>5.278586</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>5</td>\n",
       "      <td>5.330808</td>\n",
       "      <td>1</td>\n",
       "      <td>5.349438</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>6</td>\n",
       "      <td>5.403939</td>\n",
       "      <td>1</td>\n",
       "      <td>5.420289</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>7</td>\n",
       "      <td>5.470224</td>\n",
       "      <td>1</td>\n",
       "      <td>5.491140</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>8</td>\n",
       "      <td>5.572456</td>\n",
       "      <td>1</td>\n",
       "      <td>5.561991</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>9</td>\n",
       "      <td>5.646150</td>\n",
       "      <td>1</td>\n",
       "      <td>5.632842</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>10</td>\n",
       "      <td>5.685303</td>\n",
       "      <td>1</td>\n",
       "      <td>5.703693</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>11</td>\n",
       "      <td>5.731728</td>\n",
       "      <td>1</td>\n",
       "      <td>5.774544</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>12</td>\n",
       "      <td>5.844358</td>\n",
       "      <td>1</td>\n",
       "      <td>5.845395</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>13</td>\n",
       "      <td>5.917124</td>\n",
       "      <td>1</td>\n",
       "      <td>5.916246</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>14</td>\n",
       "      <td>5.962863</td>\n",
       "      <td>1</td>\n",
       "      <td>5.987097</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>15</td>\n",
       "      <td>6.007515</td>\n",
       "      <td>1</td>\n",
       "      <td>6.057948</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>16</td>\n",
       "      <td>6.231889</td>\n",
       "      <td>1</td>\n",
       "      <td>6.128799</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>17</td>\n",
       "      <td>6.229527</td>\n",
       "      <td>1</td>\n",
       "      <td>6.199650</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>18</td>\n",
       "      <td>6.268082</td>\n",
       "      <td>1</td>\n",
       "      <td>6.270501</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>19</td>\n",
       "      <td>6.211449</td>\n",
       "      <td>1</td>\n",
       "      <td>6.341352</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>20</td>\n",
       "      <td>6.310895</td>\n",
       "      <td>1</td>\n",
       "      <td>6.412203</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    educ  lwklywge  intercept      yhat\n",
       "0      0  5.022035          1  4.995182\n",
       "1      1  5.064200          1  5.066033\n",
       "2      2  5.166692          1  5.136884\n",
       "3      3  5.173006          1  5.207735\n",
       "4      4  5.264427          1  5.278586\n",
       "5      5  5.330808          1  5.349438\n",
       "6      6  5.403939          1  5.420289\n",
       "7      7  5.470224          1  5.491140\n",
       "8      8  5.572456          1  5.561991\n",
       "9      9  5.646150          1  5.632842\n",
       "10    10  5.685303          1  5.703693\n",
       "11    11  5.731728          1  5.774544\n",
       "12    12  5.844358          1  5.845395\n",
       "13    13  5.917124          1  5.916246\n",
       "14    14  5.962863          1  5.987097\n",
       "15    15  6.007515          1  6.057948\n",
       "16    16  6.231889          1  6.128799\n",
       "17    17  6.229527          1  6.199650\n",
       "18    18  6.268082          1  6.270501\n",
       "19    19  6.211449          1  6.341352\n",
       "20    20  6.310895          1  6.412203"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sumstats = educ_means.reset_index()\n",
    "sumstats['intercept'] = 1\n",
    "sumstats['yhat'] = results.predict(sumstats)\n",
    "sumstats"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f8519073358>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x7f8519073358>"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "Text(0.5, 0, 'Years of completed education')"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'Log weekly earnings, \\\\$2003')"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "Text(0.5, 0.98, 'Education ~ Wage CEF and linear fit')"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "dark"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Create plot\n",
    "plt.figure()\n",
    "educ_means.plot(style='-o')\n",
    "yhat.plot()\n",
    "plt.xlabel(\"Years of completed education\")\n",
    "plt.ylabel(\"Log weekly earnings, \\$2003\")\n",
    "plt.suptitle(\"Education ~ Wage CEF and linear fit\")\n",
    "plt.legend().set_visible(False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<ggplot: (-9223363273753235403)>"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ggplot(sumstats, aes(x = 'educ')) + geom_point(aes(y = 'lwklywge')) + geom_line(aes(y='yhat'), color = 'red') +\\\n",
    "    labs(x = 'Years of Completed Education', y = 'Log Weekly earnings, $2003', title = 'Education ~ Wage CEF and linear fit')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Propensity Score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "import patsy\n",
    "from tabulate import tabulate\n",
    "#%%"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "('nswre74.dta', <http.client.HTTPMessage at 0x7f85190733c8>)"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "('cps1re74.dta', <http.client.HTTPMessage at 0x7f8518f90828>)"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "('cps3re74.dta', <http.client.HTTPMessage at 0x7f8518f90b00>)"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#%%\n",
    "# Download data\n",
    "urllib.request.urlretrieve('http://economics.mit.edu/files/3828', 'nswre74.dta')\n",
    "urllib.request.urlretrieve('http://economics.mit.edu/files/3824', 'cps1re74.dta')\n",
    "urllib.request.urlretrieve('http://economics.mit.edu/files/3825', 'cps3re74.dta')\n",
    "#%%\n",
    "# Read the Stata files into Python\n",
    "nswre74  = pd.read_stata(\"nswre74.dta\")\n",
    "cps1re74 = pd.read_stata(\"cps1re74.dta\")\n",
    "cps3re74 = pd.read_stata(\"cps3re74.dta\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<ggplot: (8763101617612)>"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# nswre74.head()\n",
    "ggplot(nswre74, aes(x = 're78')) + geom_histogram(bins=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "#%%\n",
    "# Store list of variables for summary\n",
    "summary_vars = ['age', 'ed', 'black', 'hisp', 'nodeg', 'married', 're74', 're75']\n",
    "\n",
    "# Calculate propensity scores\n",
    "# Create formula for probit\n",
    "f = 'treat ~ ' + ' + '.join(['age', 'age2', 'ed', 'black', 'hisp', \\\n",
    "                             'nodeg', 'married', 're74', 're75'])\n",
    "#%%"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Optimization terminated successfully.\n",
      "         Current function value: 0.028323\n",
      "         Iterations 11\n"
     ]
    }
   ],
   "source": [
    "# Run probit with CPS-1\n",
    "y, X   = patsy.dmatrices(f, cps1re74, return_type = 'dataframe')\n",
    "model  = sm.Probit(y, X).fit()\n",
    "cps1re74['pscore'] = model.predict(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAkkAAAGxCAYAAAB2qSLdAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAAPYQAAD2EBqD+naQAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzt3Xt00/X9x/FXkjZtGltobYFWbIsobniZCF5aELBIFZ1WQRle2OrG3PBgHd5Af3DUqVM3vICUiWwHnCDiQD0TL1zKvFDr5oZswkFRKC2I0hYLhV5sSb+/PzjNCHzQNqT9huT5OMcjyTdJ31/eFJ8kMTgsy7IEAACAAE67BwAAAAhHRBIAAIABkQQAAGBAJAEAABgQSQAAAAZEEgAAgAGRBAAAYEAkAQAAGBBJAAAABkQSAACAAZEEAABgEGP3AMejmpqaLv+aDodDHo9HjY2Nioa/bs/tdqu5udnuMTode41M0bZXid1Gqkjba2pqaoduzzNJxwmn06mEhAQ5ndGxsri4OLtH6BLsNTJF214ldhupomWvRxMdWwYAAOggIgkAAMCASAIAADAgkgAAAAyIJAAAAAMiCQAAwIBIAgAAMCCSAAAADGz/xO3ly5drzZo12rZtm3JycnT33Xf7j02YMEF79uzxf2hXWlqaiouL/cc3bNigZ599Vl9//bUyMzN12223qU+fPgGPvXTpUjU0NGjAgAG67bbbdMIJJ0iSWlpaNG/ePL3//vtyOp3Kz8/XT3/6Uzkcji46cwAAEM5sfyYpJSVFY8eOVX5+vvH4vffeq5dfflkvv/xyQCDV1dXpkUce0ejRo7V48WINHTpUDz/8sFpaWiRJH3/8sRYvXqzp06drwYIFcjqdmjNnjv/+S5Ys0ZYtW/THP/5RTz/9tD788EO99dZbnXuyAADguGF7JOXm5urCCy9UUlJSh+5XVlam9PR05eXlKTY2VgUFBbIsS+vXr5ckrVmzRiNGjFDfvn2VkJCg8ePHq6ysTPX19ZKk1atXa9y4cerevbvS0tJ09dVXa/Xq1SE/PwAAcHyy/eW27/P000/LsixlZmbqpptuUv/+/SVJlZWVAS+tORwOZWdnq7KyUuedd54qKip07rnn+o9nZGQoJiZGO3bs0EknnaRvvvkm4P6nnHKKKisru+7EAABAWAvrSLrjjjvUt29fSVJJSYkefPBBPfPMM+rRo4caGxv97y9q4/V61djYKElqamo66vG223i93oBjzc3N8vl8crlcAferqalRTU2N/7LT6VRaWlroTrQd2mY6fLZI5XA4ouJc2Wtkira9Suw2UkXLXo8mrCOp7VkjSbr88sv1/vvv69///rdGjRolj8ejhoaGgNvX19fL4/FIkuLj4/0vrbVpaGiQx+Px36btctt93W638RfDsmXLNG/ePP/lwsJCTZo0KTQn2UEdfVnyeOZ2u+0eocuw18gUTXuV2G2kiqa9Hi6sI+lwTqdTlmVJkjIzM7VixQr/McuytG3bNo0aNUqSlJWVpfLycg0fPlyStHPnTrW0tKh3797yer1KSUnR1q1bdeKJJ0qSysvLlZmZafy6Y8aM0bBhwwLmqK2t7YxTPCqXy6WkpCTV1dXJ5/N16de2g9frPSJyIxF7jUzRtleJ3UaqSNtrcnJyh25veyT5fD75fD61traqtbVVzc3N/gipqqpSv379JB18I/bnn3/ufwYnJydHCxYs0N///ncNGTJEb775piTpnHPOkSTl5eVpxowZGjp0qDIyMrRo0SLl5OT4X2IbMWKElixZon79+qm5uVmvvfaafvzjHxtnTE1NVWpqqv9yTU1Np3xzFBUVBX3fWbNmhXAS+1mWFRW/AbVp+z6IdOw1crHbyBRtez2c7ZG0ZMkSvfTSS/7LpaWlysvL0+jRo/Xcc8/pq6++UkxMjE4++WRNnz5d6enpkg4+1Xnfffdp7ty5Ki4uVmZmpqZNm6bY2FhJ0oABA3T99dfroYceUkNDg84555yACBk3bpzq6ur061//2v85SW3PQgEAADisttev0G6Hvok7lHgm6X8SExO1b98+u8fodC6XS8nJyaqtrY2KP62x18jFbiNTpO310FeF2sP2z0kCAAAIR0QSAACAAZEEAABgQCQBAAAYEEkAAAAGRBIAAIABkQQAAGBAJAEAABgQSQAAAAZEEgAAgAGRBAAAYEAkAQAAGBBJAAAABkQSAACAAZEEAABgQCQBAAAYEEkAAAAGRBIAAIABkQQAAGBAJAEAABgQSQAAAAZEEgAAgAGRBAAAYEAkAQAAGBBJAAAABkQSAACAAZEEAABgQCQBAAAYEEkAAAAGRBIAAIABkQQAAGBAJAEAABgQSQAAAAZEEgAAgAGRBAAAYEAkAQAAGBBJAAAABkQSAACAAZEEAABgQCQBAAAYEEkAAAAGRBIAAIABkQQAAGBAJAEAABgQSQAAAAZEEgAAgAGRBAAAYEAkAQAAGBBJAAAABkQSAACAAZEEAABgQCQBAAAYxNg9wPHI7XYrLi7O7jECJCYm2j1CSMXExETcOZk4HA5JktfrlWVZNk/T+dhr5GK3kSla9no0RFIQmpub1dzcbPcYAfbt22f3CCGVmJgYcedk4nK55Ha7VV9fL5/PZ/c4nY69Ri52G5kiba8dfYKDl9sAAAAMiCQAAAADIgkAAMCASAIAADAgkgAAAAyIJAAAAAMiCQAAwIBIAgAAMCCSAAAADIgkAAAAAyIJAADAgEgCAAAwIJIAAAAMiCQAAAADIgkAAMCASAIAADAgkgAAAAyIJAAAAAMiCQAAwIBIAgAAMCCSAAAADIgkAAAAAyIJAADAgEgCAAAwIJIAAAAMiCQAAAADIgkAAMCASAIAADAgkgAAAAyIJAAAAAMiCQAAwIBIAgAAMCCSAAAADIgkAAAAAyIJAADAgEgCAAAwIJIAAAAMiCQAAAADIgkAAMCASAIAADAgkgAAAAyIJAAAAAMiCQAAwIBIAgAAMCCSAAAADIgkAAAAAyIJAADAgEgCAAAwIJIAAAAMYuweQJKWL1+uNWvWaNu2bcrJydHdd9/tP1ZRUaFnnnlG27ZtU8+ePXXLLbfoRz/6kf94aWmpnn/+eX3zzTf6wQ9+oKKiIvXo0cN/fOHChXr77bd14MABDR48WL/+9a8VGxsrSdq/f7+Ki4u1bt06eTweXXPNNSooKOi6EwcAAGErLJ5JSklJ0dixY5Wfnx9w/YEDB/TQQw/p/PPP1+LFizVu3Dg9+uij2rNnjyRp+/btmjlzpiZOnKhFixYpOztbv//97/33X7lypd59913NmDFDzz33nHbs2KFFixb5j8+dO1ctLS2aP3++HnjgAS1dulT//ve/u+akAQBAWAuLSMrNzdWFF16opKSkgOs/+eQTffvtt7r22msVGxuriy66SJmZmSotLZUkvfPOOzr33HM1YMAAxcXF6YYbblB5ebkqKyslSatXr1ZBQYF69eqlpKQkjRs3TiUlJZKkpqYmlZaWavz48UpISFB2drby8/O1atWqrj15AAAQlsIiko6msrJS2dnZcjr/N+Ypp5yiiooKSQdfiuvTp4//WEJCgnr16uU/XllZqVNOOSXgvnv37lVtba2+/PJLWZalrKws//E+ffr4AwsAAES3sHhP0tE0NjbK6/UGXOf1elVVVSXp4LNBpuONjY3G420/bmxsVFNTkxISEo5630PV1NSopqbGf9npdCotLe0Yziz0XC6X3SOElMPhiLhzMmk7x2g4V4m9RjJ2G5miZa9HE9aR5PF4VF9fH3BdfX29PB6PJCk+Pl4NDQ0BxxsaGgKOH3r/ttt6PB41NjYeEUSH3vdQy5Yt07x58/yXCwsLNWnSpGM4s9BLTk62e4SQc7vddo/QZQ5/qTmSsdfIxW4jUzTt9XBhHUmZmZlatmyZWltb/S+5lZeXa+jQoZKkrKwsbd261X/7xsZGff311/6X0DIzM1VeXq7+/ftLkrZu3apu3bopOTnZH0OVlZXKzMz0P3bbjw81ZswYDRs2zH/Z6XSqtra2E844eOE2z7Hyer1HBHIkcrlcSkpKUl1dnXw+n93jdDr2GrnYbWSKtL129AmFsIgkn88nn8+n1tZWtba2qrm5WU6nU2eddZbcbrdeeeUVFRQU6B//+IcqKio0ePBgSdLw4cN15513av369erfv79efPFFZWdn+0NnxIgRWrZsmQYOHCiv16uXXnpJI0aMkHTwWabBgwfrhRde0OTJk1VdXa2VK1fq9ttvP2K+1NRUpaam+i/X1NSE3TdHuM1zrCzLirhz+i5t3wORjr1GLnYbmaJtr4cLi0hasmSJXnrpJf/l0tJS5eXl6Te/+Y2mTZum2bNn66WXXlKPHj107733qnv37pKkk08+WUVFRSouLlZtba1OP/103XPPPf7Hyc/PV3V1te688075fD7l5ubqxhtv9B//1a9+pdmzZ6uwsFAej0djxozRwIEDu+7EAQBA2HJYlmXZPcTx5tA3cYdSUVFR0PedNWtWCCexX2Jiovbt22f3GJ3O5XIpOTlZtbW1UfGnNfYaudhtZIq0vR76qlB7hPVHAAAAANiFSAIAADAgkgAAAAyIJAAAAAMiCQAAwIBIAgAAMCCSAAAADIgkAAAAAyIJAADAgEgCAAAwIJIAAAAMiCQAAAADIgkAAMCASAIAADAgkgAAAAyIJAAAAAMiCQAAwIBIAgAAMCCSAAAADIgkAAAAAyIJAADAgEgCAAAwIJIAAAAMiCQAAAADIgkAAMCASAIAADAgkgAAAAyIJAAAAAMiCQAAwIBIAgAAMCCSAAAADIKKpLy8PH366afGY5s3b1ZeXt4xDQUAAGC3oCLpnXfeUV1dnfFYXV2d3nvvvWMaCgAAwG5Bv9zmcDiM13/wwQfq0aNH0AMBAACEg5j23vDRRx/Vo48+KulgIF188cVyOgMb69tvv9WBAwd06623hnZKAACALtbuSMrNzdWdd94py7L029/+Vtdff7169+4dcBu3260f/vCHuvLKK0M+KAAAQFdqdyQNGzZMw4YNk3TwmaRf/vKXysjI6LTBAAAA7NTuSDrU/fffH+o5AAAAwkpQkdTa2qo//elPWrp0qXbs2KGmpqaA4w6HQ1u2bAnJgAAAAHYIKpKmTJmiJ554QsOGDdPFF18st9sd6rkAAABsFVQkLVq0SA8++KCmT58e6nkAAADCQlCfk9TU1KTc3NxQzwIAABA2goqkG2+8Ua+//nqoZwEAAAgbQb3cduGFF2ratGnatWuXRo4cqe7dux9xm9GjRx/zcAAAAHYJKpLGjx8vSaqoqNCSJUuOOO5wOOTz+Y5tMgAAABsFFUnl5eWhngMAACCsBBVJWVlZoZ4DAAAgrAQVSZWVld97m8zMzGAeGgAAICwEFUnZ2dlyOBzfeRvekwQAAI5nQUXSq6++esR1tbW1WrFihT788EM99thjxzwYAACAnYKKpIKCAuP1hYWFuuOOO/Tuu+/qJz/5yTENBgAAYKegIum7XH755Ro7dqzmzJkT6ocOG263W3FxcXaPESAxMdHuEUIqJiYm4s7JpO1la6/XK8uybJ6m87HXyMVuI1O07PVoQh5JH3zwgeLj40P9sGGlublZzc3Ndo8RYN++fXaPEFKJiYkRd04mLpdLbrdb9fX1UfE+PvYaudhtZIq0vXb0CY6gIqmoqOiI65qbm7Vp0yatXbtWd911VzAPCwAAEDaCiiTT39sWHx+v3r17a86cOZowYcIxDwYAAGAnPnEbAADAwGn3AAAAAOEo6Ej6+OOPdd111yk9PV1xcXFKT0/X2LFjtX79+lDOBwAAYIugXm57//33NXLkSPXq1UvXX3+9evbsqV27dunVV19VTk6OVq1apSFDhoR6VgAAgC4TVCRNnTpVw4cP1/LlyxUT87+H+MMf/qArrrhCU6dO1dq1a0M2JAAAQFcL6uW2jz/+WEVFRQGBJB38/IiioiKtW7cuJMMBAADYJahI8nq9qqqqMh7btWuXvF7vMQ0FAABgt6Ai6corr9SUKVO0evXqgOtXr16te++9V1dddVVIhgMAALBLUO9JeuKJJ7Rx40ZdeumlSkpKUo8ePVRVVaW6ujqdd955mjFjRqjnBAAA6FJBRVJycrLKysq0fPlyrV27VrW1tUpJSdGQIUN0xRVXyOnk45cAAMDxLahIKikpUWVlpW6++eYjXlpbsGCBsrKydPHFF4dkQAAAADsE9ZTPtGnTtGvXLuOx6upqTZs27ZiGAgAAsFtQkbRx40YNGjTIeOzcc8/Vxo0bj2koAAAAuwUVSQ6HQ3v37jUeq62tlc/nO6ahAAAA7BZUJF1wwQUqLi6WZVkB11uWpTlz5uiCCy4IyXAAAAB2CeqN2w8++KAuvvhinX322SosLFR6erp27typv/zlL9q8ebPeeeedEI8JAADQtYKKpJycHJWUlOiee+7RlClT1NraKqfT6b/+wgsvDPWcAAAAXSqoSJKkwYMHq7S0VI2NjaqtrVX37t2VkJAQytkAAABsE3QktfF4PPJ4PKGYBQAAIGzw0dgAAAAGRBIAAIABkQQAAGBAJAEAABgQSQAAAAZEEgAAgAGRBAAAYEAkAQAAGBBJAAAABkQSAACAAZEEAABgQCQBAAAYEEkAAAAGRBIAAIABkQQAAGBAJAEAABgQSQAAAAZEEgAAgAGRBAAAYEAkAQAAGBBJAAAABkQSAACAAZEEAABgQCQBAAAYEEkAAAAGRBIAAIABkQQAAGBAJAEAABjE2D3A93n66af13nvvKSbmf6MWFxcrLS1NklRdXa1nnnlGmzZtUrdu3fTTn/5UQ4cO9d92w4YNevbZZ/X1118rMzNTt912m/r06eM/vnz5ci1dulQNDQ0aMGCAbrvtNp1wwgldd4IAACAsHRfPJBUUFOjll1/2/9MWSJI0Y8YMpaena+HChbr99ttVXFysiooKSVJdXZ0eeeQRjR49WosXL9bQoUP18MMPq6WlRZL08ccfa/HixZo+fboWLFggp9OpOXPm2HKOAAAgvBwXkXQ0O3fu1ObNmzV+/HjFxcXprLPO0vnnn681a9ZIksrKypSenq68vDzFxsaqoKBAlmVp/fr1kqQ1a9ZoxIgR6tu3rxISEjR+/HiVlZWpvr7eztMCAABh4LiIpBUrVuiGG25QUVGRVq1a5b++oqJCaWlpAS+P9enTx/9MUmVlZcBLaw6HQ9nZ2aqsrPTf/9DjGRkZiomJ0Y4dOzr7lAAAQJgL+/ckXXnllfr5z38ur9erjRs36vHHH5fX61Vubq6ampqOeP+Q1+tVY2OjJKmxsfE7j3/f/QEAQPQK+0jq27ev/8dnn322rrjiCpWWlio3N1fx8fFHvDTW0NAgj8cjSfJ4PGpoaAg4Xl9f7z/+ffdvU1NTo5qaGv9lp9MZ8L6ocOByueweIaQcDkfEnZNJ2zlGw7lK7DWSsdvIFC17PZqwj6TDORwOWZYlScrKylJ1dbX279/vf0Zo69atysrKkiRlZmZqxYoV/vtalqVt27Zp1KhR/vuXl5dr+PDhkg6+x6mlpUW9e/cO+JrLli3TvHnz/JcLCws1adKkTjvHYCQnJ9s9Qsi53W67R+gySUlJdo/QZdhr5GK3kSma9nq4sI+ktWvX6txzz1V8fLw+/fRTvfHGG7rlllskHXwP0amnnqqFCxfq5ptv1ueff65//vOf+v3vfy9JysnJ0YIFC/T3v/9dQ4YM0ZtvvilJOueccyRJeXl5mjFjhoYOHaqMjAwtWrRIOTk58nq9ATOMGTNGw4YN8192Op2qra3titNvt3Cb51h5vd6oeAO9y+VSUlKS6urq5PP57B6n07HXyMVuI1Ok7bWjTyg4rLanZcLU1KlTVVFRodbWVqWmpurKK6/UZZdd5j9eXV2tWbNmadOmTerevbvGjx8fEDSffPKJ5s6d6/+cpEmTJumUU07xHz/0c5LOOeccFRUVfe/nJB360lsoFRUVBX3fWbNmhXAS+yUmJmrfvn12j9HpXC6XkpOTVVtbGxW/4bLXyMVuI1Ok7TU1NbVDtw/7SApHRFLni7RvzKPhN9zIFG17ldhtpIq0vXY0ko6LjwAAAADoakQSAACAAZEEAABgQCQBAAAYEEkAAAAGRBIAAIABkQQAAGBAJAEAABgQSQAAAAZEEgAAgAGRBAAAYEAkAQAAGBBJAAAABkQSAACAAZEEAABgQCQBAAAYEEkAAAAGRBIAAIABkQQAAGBAJAEAABgQSQAAAAZEEgAAgAGRBAAAYEAkAQAAGBBJAAAABkQSAACAAZEEAABgQCQBAAAYEEkAAAAGRBIAAIABkQQAAGBAJAEAABgQSQAAAAZEEgAAgAGRBAAAYEAkAQAAGBBJAAAABkQSAACAAZEEAABgQCQBAAAYEEkAAAAGRBIAAIABkQQAAGBAJAEAABgQSQAAAAZEEgAAgAGRBAAAYEAkAQAAGBBJAAAABkQSAACAAZEEAABgQCQBAAAYxNg9wPHI7XYrLi7O7jECJCYm2j1CSMXExETcOZk4HA5JktfrlWVZNk/T+dhr5GK3kSla9no0RFIQmpub1dzcbPcYAfbt22f3CCGVmJgYcedk4nK55Ha7VV9fL5/PZ/c4nY69Ri52G5kiba8dfYKDl9sAAAAMiKTjRFVVlfbs2WP3GAAARA0i6ThRUlKit956y+4xAACIGkQSAACAAZEEAABgQCQBAAAYEEkAAAAGRBIAAIABkQQAAGBAJAEAABgQSQAAAAZEEgAAgAGRBAAAYEAkAQAAGBBJAAAABkQSAACAAZEEAABgQCQBAAAYEEkAAAAGRBIAAIABkQQAAGBAJAEAABgQSQAAAAZEEgAAgAGRBAAAYEAkAQAAGBBJAAAABkQSAACAAZEEAABgQCQBAAAYEEkAAAAGRBIAAIABkQQAAGBAJAEAABgQSQAAAAZEEgAAgAGRBAAAYEAkAQAAGBBJAAAABkQSAACAAZEEAABgQCQBAAAYEEkAAAAGMXYPgNAoKioK+r6zZs0K4SQAAEQGnkkCAAAwIJIAAAAMiCQAAAADIgkAAMCASAIAADCI+v+7bf/+/SouLta6devk8Xh0zTXXqKCgwO6xAACAzaI+kubOnauWlhbNnz9fVVVVmj59unr37q2BAwfaPVoAh8Mhy7LsHgMAgKgR1ZHU1NSk0tJSPfXUU0pISFB2drby8/O1atWqqIokPmMJAIAjRXUkffnll7IsS1lZWf7r+vTpo7KyMhunMnO73WpqatLevXvVrVs3u8fxO5bAkogsAED4iupIampqUkJCQsB1Xq9XjY2NAdfV1NSopqbGf9npdCotLa1LZmyTkpKinTt36s033wy4Pi4uTk6nU06n0z+bZVn+f7c9A+VwOORwOCTJ/++264MRqme2TjvttHY9nul4R2c49PaH/hwcfhvT9cE69Oc3mMc8fOZQzNWRxzn856O99227XWZmZsD9nU6nWltbj/h1Z1mWWltbA37dftfXM93/aL+ev2unpl8Thz5e2zzf9bVNMzgcDjmdTvl8viPOweFwGH8OjnZeR5vxaDOYzrsj3+ft/VqHa9ttZ3+dw38tBvt72LFwuVzy+Xzfe7tjOcdjOS/Tr4NgdXSv7XH66afrd7/7nU488cSQPm5niOpIio+PPyKIGhoa5PF4Aq5btmyZ5s2b579cWFioSZMmhXyeF1988ajHHn/8cf3pT3/SF198EXD9t99+G/I5gFDZs2eP3SMACDP//e9/VVRUpFNPPdXuUb5XVEfSSSedJEmqrKz0/4m3vLzc/+M2Y8aM0bBhw/yXnU6namtru25QSRMnTtSUKVNUV1fXrj/BHO+8Xq/q6+vtHqPTuVwuJSUlsdcIE217ldhtpOrMvXb1f0clKTk5uUO3j+pIio+P1+DBg/XCCy9o8uTJqq6u1sqVK3X77bcH3C41NVWpqan+yzU1NbZ9c/h8vqj4xrQsKyrOsw17jUzRsleJ3UaqaNvr4aI6kiTpV7/6lWbPnq3CwkJ5PB6NGTMm7P7PNgAA0PWiPpJOOOEETZ061e4xAABAmOGvJQEAADAgkgAAAAyIJAAAAAMiCQAAwIBIAgAAMCCSAAAADIgkAAAAAyIJAADAgEgCAAAwIJIAAAAMiCQAAAADIgkAAMDAYVmWZfcQ+H41NTVatmyZxowZo9TUVLvHQYiw18jEXiMXu40uPJN0nKipqdG8efNUU1Nj9ygIIfYamdhr5GK30YVIAgAAMCCSAAAADFwPPPDAA3YPgfbxeDwaNGiQEhIS7B4FIcReIxN7jVzsNnrwxm0AAAADXm4DAAAwIJIAAAAMYuweAN9v//79Ki4u1rp16+TxeHTNNdeooKDA7rHQTu3d36effqrFixfriy++kCSdfvrpmjBhgjIyMrp6ZLRDMN+XJSUlmjlzpiZOnKhRo0Z10aToqI7strm5Wc8//7zee+89NTc3KyMjQ4888gjvV4oQRNJxYO7cuWppadH8+fNVVVWl6dOnq3fv3ho4cKDdo6Ed2ru/+vp6XXLJJbrnnnvkdru1aNEiPfzww5ozZ45Nk+O7dPT7sq6uTkuXLlVmZmYXT4qO6shu58yZo6amJs2aNUvdunVTRUWFYmNjbZganYGX28JcU1OTSktLNX78eCUkJCg7O1v5+flatWqV3aOhHTqyv4EDB+qiiy6S1+tVbGysrr76au3YsUN1dXU2TI7vEsz35fz581VQUKCkpKQunBQd1ZHd7tixQ2VlZZo0aZKSk5PldDrVp08fIimCEElh7ssvv5RlWcrKyvJf16dPH1VWVto4FdrrWPa3YcMGJScn8x/VMNTRvW7YsEHbt29Xfn5+V42IIHVkt59//rl69OihJUuW6KabbtKtt96qlStXduW46GREUphramo64rVtr9erxsZGmyZCRwS7v6+//lpz587VhAkTOnM8BKkje21padGzzz6riRMnyunkt9xw15HdVldXq6KiQm63W/Pnz9fkyZM1f/58bdiwoavGRSfjOzbMxcfHH/HN2dDQII/HY9NE6Ihg9lddXa3p06drzJgxuuiiizp7RAShI3t95ZVXdOaZZ6pv375dNR6OQUd2GxcXJ6fMsgTkAAAG1klEQVTTqXHjxik2NlannXaaBg8erI8++qirxkUn443bYe6kk06SJFVWVvrf8FleXs6bP48THd1fTU2Npk2bpksvvVRXX311l82JjunIXv/zn/+ooqJCH3zwgaSD/+fU1q1btXnzZt1+++1dNzTapSO7zc7O7srRYAOeSQpz8fHxGjx4sF544QU1NDSooqJCK1eu1MiRI+0eDe3Qkf3t3r1b//d//6fhw4fr2muvtWFatFdH9nrvvfdq9uzZmjlzpmbOnKlTTz1V1113nX7xi1/YMDm+T0d2e+aZZ6pXr17661//Kp/Ppy1btqi0tFTnnXeeDZOjM/DXkhwH9u/fr9mzZ/s/s2P06NF8TtJx5Lv2N3bsWN1///0644wztHjxYi1evFjx8fEB9y8uLlZaWpodo+M7tHevh7vvvvt00UUX8TlJYawju92xY4dmz56tLVu2KCUlRdddd50uueQSO8dHCBFJAAAABrzcBgAAYEAkAQAAGBBJAAAABkQSAACAAZEEAABgQCQBAAAYEEkAAAAGRBIAAIABkQQAAGBAJAEAABgQSQDQDo2NjXaPAKCLEUkAwkZhYaHOPPNMvfXWWzrzzDMVHx+vgQMH6sMPP/Tf5m9/+5sGDRqkE044Qd27d9egQYP05ptvBjzOX/7yFw0YMEDx8fFKTU3V5ZdfroqKCv/xTz75RJdeeqm8Xq+6deuma6+9VpWVlQGP4XA49Nhjj2nKlCnq1auXevTo4T9WVlamvLw8//1vuOEGVVVVddLPCgC7EEkAwspXX32lW2+9VXfffbdefvllxcXF6dJLL1VVVZW2bNmia6+9VmeccYZeffVVLVmyRGPHjlVtba3//n/4wx/0s5/9TAMHDtQrr7yiP//5zzrttNNUXV0tSdq+fbuGDh2q3bt3a+HChXr22We1bt06DRs2TPv27QuYZebMmdq8ebP+/Oc/a+HChZIOBtLw4cPVrVs3LVmyRM8995w++ugj/98SDyCCWAAQJn72s59ZkqySkhL/dXv27LESExOtqVOnWn/9618tSVZdXZ3x/nv27LESEhKsW2655ahfY/LkyZbX67V2797tv27Tpk2Ww+GwZs2a5b9OktW/f3+rtbU14P5Dhw61cnNzA67fuHGj5XA4rDfeeKPD5wwgfPFMEoCw0q1bN+Xl5QVcvuSSS/SPf/xDZ599tlwul2644Qa9/vrr2rt3b8B9y8rK1NDQoF/84hdHffz3339feXl5SklJ8V/3gx/8QD/60Y+0du3agNuOGjVKDofDf7mhoUGlpaW67rrr5PP5dODAAR04cED9+vXTySefrI8++uhYTx9AGCGSAISVtLS0I67r2bOnvvrqK/Xr10/Lly/X3r17dc011ygtLU1XXXWV//1Eu3fvliRlZGQc9fFra2vVs2dP49f45ptvjrju8Pv6fD5NnjxZsbGxAf9UVlZq+/btHT5fAOErxu4BAOBQbe8dOtSuXbuUnp4uSbrssst02WWXqa6uTm+//bYmT56sm2++WSUlJTrxxBMlSTt37lTv3r2Nj5+SkmJ8k/WuXbvUr1+/gOsOfRZJkrp37y6Hw6H77rtPV1999RGPkZqa2r6TBHBc4JkkAGFl7969WrNmTcDl1atX64ILLgi4XVJSksaOHatx48Zp06ZNkqScnBwlJCRo/vz5R338IUOGqKSkJODN3p999pn++9//asiQId85m9frVU5OjjZt2qRBgwYd8U92dnYQZwwgXLkeeOCBB+weAgAk6bXXXtP27du1YsUKpaSkqKKiQhMnTtSuXbv04osvatGiRZo9e7aampq0e/duvfPOO3ryySc1cuRIjR49WvHx8YqNjdVjjz2mr776Si6XS5999pnmzp2rE088URkZGerfv7+ee+45vfHGG+rRo4fWr1+vCRMmKCkpSXPnzlVcXJwk6cEHH1R+fr5yc3MDZvzhD3+oadOm6ZNPPlFMTIx27dqltWvX6oknnlBycjKhBEQQXm4DEFbS09P1+OOP6+6779aWLVt0xhlnaMWKFerZs6fOPvtsvf7667rjjju0e/du9erVS9dff70eeugh//3vuecepaWl6amnntKCBQuUmJionJwc/+ccnXzyyXr33Xd111136cYbb5TL5dLIkSP15JNPKjEx8Xvny83N1dq1a3X//ffr5ptvVnNzs3r37q0RI0bo1FNP7bSfFwBdz2FZlmX3EAAgHfwwyX/961/asGGD3aMAAO9JAgAAMCGSAAAADHi5DQAAwIBnkgAAAAyIJAAAAAMiCQAAwIBIAgAAMCCSAAAADIgkAAAAAyIJAADAgEgCAAAwIJIAAAAM/h+vPmFZBfrX5wAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<ggplot: (-9223363273753158382)>"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# nswre74.head()\n",
    "ggplot(cps1re74, aes(x = 'pscore')) + geom_histogram(bins=30) + geom_density()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Optimization terminated successfully.\n",
      "         Current function value: 0.355062\n",
      "         Iterations 7\n"
     ]
    }
   ],
   "source": [
    "#%%\n",
    "# Run probit with CPS-3\n",
    "y, X   = patsy.dmatrices(f, cps3re74, return_type = 'dataframe')\n",
    "model  = sm.Probit(y, X).fit()\n",
    "cps3re74['pscore'] = model.predict(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<ggplot: (8763101187788)>"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# nswre74.head()\n",
    "ggplot(cps3re74, aes(x = 'pscore')) + geom_histogram(bins=30) + geom_density()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "#%%\n",
    "# Create function to summarize data\n",
    "def summarize(dataset, conditions):\n",
    "  stats          = dataset[summary_vars][conditions].mean()\n",
    "  stats['count'] = sum(conditions)\n",
    "  return stats\n",
    "#%%\n",
    "# Summarize data\n",
    "nswre74_treat_stats    = summarize(nswre74, nswre74.treat == 1)\n",
    "nswre74_control_stats  = summarize(nswre74, nswre74.treat == 0)\n",
    "cps1re74_control_stats = summarize(cps1re74, cps1re74.treat == 0)\n",
    "cps3re74_control_stats = summarize(cps3re74, cps3re74.treat == 0)\n",
    "cps1re74_ptrim_stats   = summarize(cps1re74, (cps1re74.treat == 0)   & \\\n",
    "                                             (cps1re74.pscore > 0.1) & \\\n",
    "                                             (cps1re74.pscore < 0.9))\n",
    "cps3re74_ptrim_stats   = summarize(cps3re74, (cps3re74.treat == 0)   & \\\n",
    "                                             (cps3re74.pscore > 0.1) & \\\n",
    "                                             (cps3re74.pscore < 0.9))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "|         |    NSW Treat |   NSW Control |    Full CPS-1 |   Full CPS-3 |   P-score CPS-1 |   P-score CPS-3 |\n",
      "|:--------|-------------:|--------------:|--------------:|-------------:|----------------:|----------------:|\n",
      "| age     |   25.8162    |     25.0538   |    33.2252    |    28.0303   |      25.6278    |       25.9682   |\n",
      "| ed      |   10.3459    |     10.0885   |    12.0275    |    10.2354   |      10.4858    |       10.4204   |\n",
      "| black   |    0.843243  |      0.826923 |     0.0735368 |     0.202797 |       0.957386  |        0.515924 |\n",
      "| hisp    |    0.0594595 |      0.107692 |     0.072036  |     0.142191 |       0.0284091 |        0.203822 |\n",
      "| nodeg   |    0.708108  |      0.834615 |     0.295835  |     0.596737 |       0.599432  |        0.630573 |\n",
      "| married |    0.189189  |      0.153846 |     0.711731  |     0.512821 |       0.261364  |        0.286624 |\n",
      "| re74    | 2095.57      |   2107.03     | 14016.4       |  5619.23     |    2821.37      |     2968.51     |\n",
      "| re75    | 1532.06      |   1266.91     | 13650.9       |  2466.49     |    1949.81      |     1858.79     |\n",
      "| count   |  185         |    260        | 15992         |   429        |     352         |      157        |\n"
     ]
    }
   ],
   "source": [
    "#%%\n",
    "# Combine summary stats, add header and print to markdown\n",
    "frames = [nswre74_treat_stats,\n",
    "          nswre74_control_stats,\n",
    "          cps1re74_control_stats,\n",
    "          cps3re74_control_stats,\n",
    "          cps1re74_ptrim_stats,\n",
    "          cps3re74_ptrim_stats]\n",
    "#%%\n",
    "summary_stats = pd.concat(frames, axis = 1)\n",
    "header        = [\"NSW Treat\", \"NSW Control\", \\\n",
    "                 \"Full CPS-1\", \"Full CPS-3\", \\\n",
    "                 \"P-score CPS-1\", \"P-score CPS-3\"]\n",
    "\n",
    "print(tabulate(summary_stats, header, tablefmt = \"pipe\"))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Instrumental variables"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "from matplotlib.ticker import FormatStrFormatter"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Angrist and Krueger"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Calculate means by educ and lwklywge\n",
    "groupbybirth = pums.groupby(['yob', 'qob'])\n",
    "birth_means  = groupbybirth['lwklywge', 'educ'].mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create function to plot figures\n",
    "def plot_qob(yvar, ax, title, ylabel):\n",
    "    values = yvar.values\n",
    "    ax.plot(values, color = 'k')\n",
    "\n",
    "    for i, y in enumerate(yvar):\n",
    "        qob = yvar.index.get_level_values('qob')[i]\n",
    "        ax.annotate(qob,\n",
    "                    (i, y),\n",
    "                    xytext = (-5, 5),\n",
    "                    textcoords = 'offset points')\n",
    "        if qob == 1:\n",
    "            ax.scatter(i, y, marker = 's', facecolors = 'none', edgecolors = 'y')\n",
    "        else:\n",
    "            ax.scatter(i, y, marker = 's', color = 'r')\n",
    "\n",
    "    ax.set_xticks(range(0, len(yvar), 4))\n",
    "    ax.set_xticklabels(yvar.index.get_level_values('yob')[1::4])\n",
    "    ax.set_title(title)\n",
    "    ax.set_ylabel(ylabel)\n",
    "    ax.yaxis.set_major_formatter(FormatStrFormatter('%.2f'))\n",
    "    ax.set_xlabel(\"Year of birth\")\n",
    "    ax.margins(0.1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x648 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "dark"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1, ax2) = plt.subplots(2, figsize=(12, 9), sharex = True)\n",
    "\n",
    "plot_qob(yvar   = birth_means['educ'],\n",
    "         ax     = ax1,\n",
    "         title  = 'A. Average education by quarter of birth (first stage)',\n",
    "         ylabel = 'Years of education')\n",
    "\n",
    "plot_qob(yvar   = birth_means['lwklywge'],\n",
    "         ax     = ax2,\n",
    "         title  = 'B. Average weekly wage by quarter of birth (reduced form)',\n",
    "         ylabel = 'Log weekly earnings')\n",
    "\n",
    "fig.tight_layout()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### IV estimation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "from linearmodels.iv import IV2SLS"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "qobs = pd.get_dummies(pums.qob, prefix='qob')\n",
    "pums = pd.concat([pums, qobs], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>lwklywge</th>\n",
       "      <th>educ</th>\n",
       "      <th>yob</th>\n",
       "      <th>qob</th>\n",
       "      <th>pob</th>\n",
       "      <th>qob_1</th>\n",
       "      <th>qob_2</th>\n",
       "      <th>qob_3</th>\n",
       "      <th>qob_4</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5.790019</td>\n",
       "      <td>12</td>\n",
       "      <td>30</td>\n",
       "      <td>1</td>\n",
       "      <td>45</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5.952494</td>\n",
       "      <td>11</td>\n",
       "      <td>30</td>\n",
       "      <td>1</td>\n",
       "      <td>45</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>5.315949</td>\n",
       "      <td>12</td>\n",
       "      <td>30</td>\n",
       "      <td>1</td>\n",
       "      <td>45</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>5.595926</td>\n",
       "      <td>12</td>\n",
       "      <td>30</td>\n",
       "      <td>1</td>\n",
       "      <td>45</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>6.068915</td>\n",
       "      <td>12</td>\n",
       "      <td>30</td>\n",
       "      <td>1</td>\n",
       "      <td>37</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   lwklywge  educ  yob  qob  pob  qob_1  qob_2  qob_3  qob_4\n",
       "0  5.790019    12   30    1   45      1      0      0      0\n",
       "1  5.952494    11   30    1   45      1      0      0      0\n",
       "2  5.315949    12   30    1   45      1      0      0      0\n",
       "3  5.595926    12   30    1   45      1      0      0      0\n",
       "4  6.068915    12   30    1   37      1      0      0      0"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pums.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "formula = 'lwklywge ~ 1 [educ ~ qob_2 + qob_3 + qob_4]'\n",
    "mod = IV2SLS.from_formula(formula, pums).fit(cov_type='robust')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                          IV-2SLS Estimation Summary                          \n",
      "==============================================================================\n",
      "Dep. Variable:               lwklywge   R-squared:                      0.0937\n",
      "Estimator:                    IV-2SLS   Adj. R-squared:                 0.0937\n",
      "No. Observations:              329509   F-statistic:                    27.603\n",
      "Date:                Fri, Jan 04 2019   P-value (F-stat)                0.0000\n",
      "Time:                        23:09:03   Distribution:                  chi2(1)\n",
      "Cov. Estimator:                robust                                         \n",
      "                                                                              \n",
      "                             Parameter Estimates                              \n",
      "==============================================================================\n",
      "            Parameter  Std. Err.     T-stat    P-value    Lower CI    Upper CI\n",
      "------------------------------------------------------------------------------\n",
      "Intercept      4.5898     0.2494     18.404     0.0000      4.1010      5.0786\n",
      "educ           0.1026     0.0195     5.2539     0.0000      0.0643      0.1409\n",
      "==============================================================================\n",
      "\n",
      "Endogenous: educ\n",
      "Instruments: qob_2, qob_3, qob_4\n",
      "Robust Covariance (Heteroskedastic)\n",
      "Debiased: False\n"
     ]
    }
   ],
   "source": [
    "print(mod.summary)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Clustering / Fixed Effects "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "import statsmodels.stats.sandwich_covariance as sw"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:               lwklywge   R-squared:                       0.117\n",
      "Model:                            OLS   Adj. R-squared:                  0.117\n",
      "Method:                 Least Squares   F-statistic:                 3.458e+04\n",
      "Date:                Fri, 04 Jan 2019   Prob (F-statistic):               0.00\n",
      "Time:                        23:30:54   Log-Likelihood:            -3.1935e+05\n",
      "No. Observations:              329509   AIC:                         6.387e+05\n",
      "Df Residuals:                  329507   BIC:                         6.387e+05\n",
      "Df Model:                           1                                         \n",
      "Covariance Type:                  HC1                                         \n",
      "==============================================================================\n",
      "                 coef    std err          z      P>|z|      [0.025      0.975]\n",
      "------------------------------------------------------------------------------\n",
      "Intercept      4.9952      0.005    984.491      0.000       4.985       5.005\n",
      "educ           0.0709      0.000    185.949      0.000       0.070       0.072\n",
      "==============================================================================\n",
      "Omnibus:                   191064.440   Durbin-Watson:                   1.870\n",
      "Prob(Omnibus):                  0.000   Jarque-Bera (JB):          4082110.366\n",
      "Skew:                          -2.377   Prob(JB):                         0.00\n",
      "Kurtosis:                      19.575   Cond. No.                         53.3\n",
      "==============================================================================\n",
      "\n",
      "Warnings:\n",
      "[1] Standard Errors are heteroscedasticity robust (HC1)\n"
     ]
    }
   ],
   "source": [
    "base_mincer = smf.ols(formula='lwklywge ~ educ', data=pums).fit(cov_type='HC1')\n",
    "print(base_mincer.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:               lwklywge   R-squared:                       0.117\n",
      "Model:                            OLS   Adj. R-squared:                  0.117\n",
      "Method:                 Least Squares   F-statistic:                     1610.\n",
      "Date:                Fri, 04 Jan 2019   Prob (F-statistic):           1.07e-39\n",
      "Time:                        23:30:55   Log-Likelihood:            -3.1935e+05\n",
      "No. Observations:              329509   AIC:                         6.387e+05\n",
      "Df Residuals:                  329507   BIC:                         6.387e+05\n",
      "Df Model:                           1                                         \n",
      "Covariance Type:              cluster                                         \n",
      "==============================================================================\n",
      "                 coef    std err          t      P>|t|      [0.025      0.975]\n",
      "------------------------------------------------------------------------------\n",
      "Intercept      4.9952      0.031    162.935      0.000       4.934       5.057\n",
      "educ           0.0709      0.002     40.126      0.000       0.067       0.074\n",
      "==============================================================================\n",
      "Omnibus:                   191064.440   Durbin-Watson:                   1.870\n",
      "Prob(Omnibus):                  0.000   Jarque-Bera (JB):          4082110.366\n",
      "Skew:                          -2.377   Prob(JB):                         0.00\n",
      "Kurtosis:                      19.575   Cond. No.                         53.3\n",
      "==============================================================================\n",
      "\n",
      "Warnings:\n",
      "[1] Standard Errors are robust tocluster correlation (cluster)\n"
     ]
    }
   ],
   "source": [
    "cluster_mincer = smf.ols(formula='lwklywge ~ educ', data=pums).fit(cov_type='cluster',\n",
    "                                                        cov_kwds={'groups': pums['pob']},\n",
    "                                                        use_t=True)\n",
    "print(cluster_mincer.summary())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                            OLS Regression Results                            \n",
      "==============================================================================\n",
      "Dep. Variable:               lwklywge   R-squared:                       0.128\n",
      "Model:                            OLS   Adj. R-squared:                  0.128\n",
      "Method:                 Least Squares   F-statistic:                     759.0\n",
      "Date:                Fri, 04 Jan 2019   Prob (F-statistic):               0.00\n",
      "Time:                        23:32:30   Log-Likelihood:            -3.1726e+05\n",
      "No. Observations:              329509   AIC:                         6.346e+05\n",
      "Df Residuals:                  329457   BIC:                         6.352e+05\n",
      "Df Model:                          51                                         \n",
      "Covariance Type:                  HC1                                         \n",
      "================================================================================\n",
      "                   coef    std err          t      P>|t|      [0.025      0.975]\n",
      "--------------------------------------------------------------------------------\n",
      "Intercept        4.9322      0.009    547.748      0.000       4.915       4.950\n",
      "C(pob)[T.2]      0.2584      0.095      2.720      0.007       0.072       0.445\n",
      "C(pob)[T.4]      0.1459      0.019      7.526      0.000       0.108       0.184\n",
      "C(pob)[T.5]      0.0457      0.012      3.927      0.000       0.023       0.069\n",
      "C(pob)[T.6]      0.1729      0.010     17.644      0.000       0.154       0.192\n",
      "C(pob)[T.8]      0.1214      0.014      8.583      0.000       0.094       0.149\n",
      "C(pob)[T.9]      0.1352      0.012     10.965      0.000       0.111       0.159\n",
      "C(pob)[T.10]     0.0888      0.024      3.693      0.000       0.042       0.136\n",
      "C(pob)[T.11]     0.1534      0.019      8.128      0.000       0.116       0.190\n",
      "C(pob)[T.12]     0.0013      0.013      0.097      0.923      -0.024       0.026\n",
      "C(pob)[T.13]    -0.0196      0.011     -1.845      0.065      -0.041       0.001\n",
      "C(pob)[T.15]     0.1531      0.042      3.637      0.000       0.071       0.236\n",
      "C(pob)[T.16]     0.1043      0.017      5.989      0.000       0.070       0.138\n",
      "C(pob)[T.17]     0.2064      0.009     22.922      0.000       0.189       0.224\n",
      "C(pob)[T.18]     0.1533      0.010     15.276      0.000       0.134       0.173\n",
      "C(pob)[T.19]     0.0994      0.011      8.820      0.000       0.077       0.122\n",
      "C(pob)[T.20]     0.0748      0.012      6.186      0.000       0.051       0.098\n",
      "C(pob)[T.21]     0.1115      0.010     10.653      0.000       0.091       0.132\n",
      "C(pob)[T.22]     0.0947      0.012      8.109      0.000       0.072       0.118\n",
      "C(pob)[T.23]    -0.0111      0.014     -0.778      0.437      -0.039       0.017\n",
      "C(pob)[T.24]     0.1178      0.013      9.415      0.000       0.093       0.142\n",
      "C(pob)[T.25]     0.1112      0.010     11.348      0.000       0.092       0.130\n",
      "C(pob)[T.26]     0.2268      0.009     24.538      0.000       0.209       0.245\n",
      "C(pob)[T.27]     0.1501      0.011     13.860      0.000       0.129       0.171\n",
      "C(pob)[T.28]    -0.0215      0.012     -1.801      0.072      -0.045       0.002\n",
      "C(pob)[T.29]     0.1138      0.010     11.017      0.000       0.094       0.134\n",
      "C(pob)[T.30]     0.0775      0.020      3.900      0.000       0.039       0.116\n",
      "C(pob)[T.31]     0.0922      0.014      6.726      0.000       0.065       0.119\n",
      "C(pob)[T.32]     0.1524      0.041      3.710      0.000       0.072       0.233\n",
      "C(pob)[T.33]     0.0203      0.019      1.092      0.275      -0.016       0.057\n",
      "C(pob)[T.34]     0.1712      0.010     16.898      0.000       0.151       0.191\n",
      "C(pob)[T.35]     0.0703      0.018      3.975      0.000       0.036       0.105\n",
      "C(pob)[T.36]     0.1623      0.009     18.981      0.000       0.146       0.179\n",
      "C(pob)[T.37]    -0.0566      0.010     -5.664      0.000      -0.076      -0.037\n",
      "C(pob)[T.38]     0.1305      0.017      7.800      0.000       0.098       0.163\n",
      "C(pob)[T.39]     0.1654      0.009     18.398      0.000       0.148       0.183\n",
      "C(pob)[T.40]     0.0924      0.011      8.504      0.000       0.071       0.114\n",
      "C(pob)[T.41]     0.1408      0.015      9.255      0.000       0.111       0.171\n",
      "C(pob)[T.42]     0.1316      0.009     15.441      0.000       0.115       0.148\n",
      "C(pob)[T.44]     0.0463      0.017      2.727      0.006       0.013       0.080\n",
      "C(pob)[T.45]    -0.0805      0.012     -6.496      0.000      -0.105      -0.056\n",
      "C(pob)[T.46]     0.0785      0.017      4.604      0.000       0.045       0.112\n",
      "C(pob)[T.47]     0.0480      0.011      4.533      0.000       0.027       0.069\n",
      "C(pob)[T.48]     0.0882      0.009      9.445      0.000       0.070       0.107\n",
      "C(pob)[T.49]     0.1411      0.015      9.251      0.000       0.111       0.171\n",
      "C(pob)[T.50]    -0.0375      0.021     -1.779      0.075      -0.079       0.004\n",
      "C(pob)[T.51]     0.0319      0.011      2.910      0.004       0.010       0.053\n",
      "C(pob)[T.53]     0.1877      0.013     14.544      0.000       0.162       0.213\n",
      "C(pob)[T.54]     0.1274      0.011     11.757      0.000       0.106       0.149\n",
      "C(pob)[T.55]     0.1281      0.010     12.378      0.000       0.108       0.148\n",
      "C(pob)[T.56]     0.1438      0.026      5.522      0.000       0.093       0.195\n",
      "educ             0.0671      0.000    173.051      0.000       0.066       0.068\n",
      "==============================================================================\n",
      "Omnibus:                   193121.979   Durbin-Watson:                   1.903\n",
      "Prob(Omnibus):                  0.000   Jarque-Bera (JB):          4260317.292\n",
      "Skew:                          -2.403   Prob(JB):                         0.00\n",
      "Kurtosis:                      19.947   Cond. No.                         867.\n",
      "==============================================================================\n",
      "\n",
      "Warnings:\n",
      "[1] Standard Errors are heteroscedasticity robust (HC1)\n"
     ]
    }
   ],
   "source": [
    "fe_mincer = smf.ols(formula='lwklywge ~ educ + C(pob)', data=pums).fit(cov_type='HC1',\n",
    "                                                        use_t=True)\n",
    "print(fe_mincer.summary())"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}