#!/usr/bin/env python
# coding: utf-8

# # Unemployment insurance vs. UBI
# 
# Compare the poverty and distributional impacts of giving \$600 per unemployment insurance beneficiary--as proposed in the latest CARES Act--to using the same amount for a UBI.
# 
# Data: 2018 ASEC
# 
# ## Setup

# In[1]:


import pandas as pd
import numpy as np
import microdf as mdf
import matplotlib.pyplot as plt
import matplotlib as mpl
import os


# ## Load data

# In[2]:


ASEC_F = '~/data/asec/2019/pppub19.csv'

SPM_COLS = ['povthreshold', 'resources', 'poor', 'numper', 'numkids',
            'numadults', 'id', 'weight']
OTHER_COLS = ['A_AGE', 'MARSUPWT', 'UC_VAL']
cols = ['SPM_' + i.upper() for i in SPM_COLS] + OTHER_COLS

raw = pd.read_csv(ASEC_F, usecols=cols)


# ## Preprocess

# In[3]:


df = raw.copy(deep=True)


# In[4]:


df.columns = map(str.lower, df.columns)


# Add true weight by dividing by 100.

# In[5]:


df['w'] = df.marsupwt / 100.
df['spm_w'] = df.spm_weight / 100.


# In[6]:


mdf.weighted_sum(df, 'uc_val', 'w') / 1e9


# ## Simulate

# In[7]:


UI_BONUS = 600 * 17


# In[8]:


df['ui_bonus'] = np.where(df.uc_val > 0, UI_BONUS, 0)


# In[9]:


spmu = df.groupby('spm_id')[['ui_bonus']].sum()
spmu_raw = df[['spm_resources', 'spm_povthreshold', 'spm_id', 'spm_numper',
               'spm_w']].drop_duplicates().set_index('spm_id')
spmu = spmu.join(spmu_raw)
spmu


# In[10]:


budget = mdf.weighted_sum(spmu, 'ui_bonus', 'spm_w')
budget / 1e9


# In[11]:


ubi = budget / (spmu.spm_numper * spmu.spm_w).sum()
ubi


# In[12]:


spmu['ubi'] = ubi * spmu.spm_numper


# In[13]:


mdf.weighted_sum(spmu, 'ubi', 'spm_w') - budget


# Add new resource columns and merge back to person level dataset.

# In[14]:


spmu['spm_resources_ui'] = spmu.spm_resources + spmu.ui_bonus
spmu['spm_resources_ubi'] = spmu.spm_resources + spmu.ubi


# In[15]:


df = df.merge(spmu[['spm_resources_ui', 'spm_resources_ubi']], on='spm_id')


# ## Calculate effects

# In[16]:


df['spm_poor_ui'] = np.where(df.spm_resources_ui < df.spm_povthreshold, 1, 0)
df['spm_poor_ubi'] = np.where(df.spm_resources_ubi < df.spm_povthreshold, 1,
                              0)


# In[17]:


pov_rate_base = mdf.weighted_mean(df, 'spm_poor', 'w')
pov_rate_ui = mdf.weighted_mean(df, 'spm_poor_ui', 'w')
pov_rate_ubi = mdf.weighted_mean(df, 'spm_poor_ubi', 'w')


# In[18]:


(pov_rate_base - pov_rate_ubi) / (pov_rate_base - pov_rate_ui)


# Poverty gap.

# In[19]:


def pov_gap(resource_col):
    pov_gap = np.maximum(spmu.spm_povthreshold - spmu[resource_col], 0)
    return (pov_gap * spmu.spm_w).sum() / 1e9


# In[20]:


pov_gap_base = pov_gap('spm_resources')
pov_gap_ui = pov_gap('spm_resources_ui')
pov_gap_ubi = pov_gap('spm_resources_ubi')


# In[21]:


(pov_gap_base - pov_gap_ubi) / (pov_gap_base - pov_gap_ui)


# ## Poverty rate of those on UI

# In[22]:


def pov_rate(df):
    return mdf.weighted_mean(df, 'spm_poor', 'w')


# In[23]:


pov_rate(df[df.uc_val > 0])


# In[24]:


pov_rate(df[df.uc_val == 0])


# ## Where in the wage distribution are people who collect UI?
# 
# [Pelosi said:](https://www.cnbc.com/2020/03/25/bernie-sanders-4-gop-senators-threaten-to-hold-up-coronavirus-stimulus-bill.html)
# >Please don’t resent our lowest paid workers in America for getting \$600.

# In[25]:


spmu['pov_ratio'] = spmu.spm_resources / spmu.spm_povthreshold


# In[26]:


spmu_ui = spmu[spmu.ui_bonus > 0].copy(deep=True)
spmu_noui = spmu[spmu.ui_bonus == 0].copy(deep=True)


# In[27]:


mdf.weighted_median(spmu, 'pov_ratio', 'spm_w')


# In[28]:


mdf.weighted_median(spmu_ui, 'pov_ratio', 'spm_w')


# In[29]:


mdf.weighted_median(spmu_noui, 'pov_ratio', 'spm_w')


# ### t test
# 
# Highly significant (2nd element is p value).

# In[30]:


from statsmodels.stats import weightstats


# In[31]:


weightstats.ttest_ind(spmu_ui.pov_ratio, spmu_noui.pov_ratio, 
                      weights=(spmu_ui.spm_w, spmu_noui.spm_w))