def __init__(self,
*,
base,
targ_mode,
compare=None,
setup=None,
base_name='Model',
compare_name='Data',
graphs_title_add=None,
moments_aux=None,
noplot=False):
if setup is None: self.setup = ModelSetup()
if type(base) is str:
self.moments = filer(base, 0, 0, repeat=False)
else:
self.moments = base
self.targ_mode = targ_mode
self.base_name = base_name
self.compare_name = compare_name
if graphs_title_add:
self.graphs_title_add = '\n ' + graphs_title_add
else:
self.graphs_title_add = ''
if type(compare) is str:
targ_load = filer(compare, 0, 0, repeat=False)
# emulating targets
self.targets = {key: (targ_load[key], 0.0) for key in targ_load}
else:
self.targets = all_targets(targ_mode)
if not noplot:
try:
self.print_things()
except:
print('failed to print')
self.moments_aux = moments_aux
if not noplot: self.plot()
def __init__(self, iterator_name='default', verbose=False, **kwargs):
self.mstart = self.get_mem()
self.mlast = self.get_mem()
self.verbose = verbose
self.setup = ModelSetup(**kwargs)
self.dtype = self.setup.dtype
self.iterator, self.initializer = self._get_iterator(iterator_name)
self.start = default_timer()
self.last = default_timer()
self.time_dict = dict()
self.solve()
def __init__(self, verbose=False, **kwargs):
self.mstart = self.get_mem()
self.mlast = self.get_mem()
self.verbose = verbose
self.start = default_timer()
self.last = default_timer()
self.time_dict = dict()
self.gpu = gpu
self.setup = ModelSetup(**kwargs)
self.dtype = self.setup.dtype
self.time('Creation of setup')
self.solve()
def __init__(self,iterator_name='default-timed',verbose=False,
solve_till=None,display_v=False,draw=False,**kwargs):
self.mstart = self.get_mem()
self.mlast = self.get_mem()
self.verbose = verbose
self.setup = ModelSetup(**kwargs)
self.dtype = self.setup.dtype
self.iterator, self.initializer = self._get_iterator(iterator_name)
self.start = default_timer()
self.last = default_timer()
self.time_dict = dict()
self.display_v = display_v
if solve_till is not None:
T = self.setup.pars['T']
if solve_till < 0: solve_till = T + solve_till
print('T is {}, but solving till T = {}'.format(T,solve_till))
self.solve(till=solve_till,draw=draw)
def __init__(self,*,base,targ_mode,compare=None,
setup=None,
base_name='Model',
compare_name='Data',
graphs_title_add=None,
moments_aux=None):
if setup is None: self.setup = ModelSetup()
if type(base) is str:
self.moments = filer(base,0,0)
else:
self.moments = base
self.targ_mode = targ_mode
self.base_name = base_name
self.compare_name = compare_name
if graphs_title_add:
self.graphs_title_add = '\n ' + graphs_title_add
else:
self.graphs_title_add = ''
if type(compare) is str:
targ_load = filer(compare,0,0)
# emulating targets
self.targets = {key: (targ_load[key],0.0) for key in targ_load}
else:
self.targets = all_targets(targ_mode)
try:
self.print_things()
except:
print('failed to print')
try:
self.plot_estimates()
except:
print('failed to plot estimates')
try:
self.plot_hazards()
except:
print('failed to plot hazards')
try:
self.plot_cumulative()
except:
print('failed to plot cumulative')
try:
self.plot_by_years_after_marriage()
except:
print('failed to plot by years after marriage')
try:
self.plot_kfmf()
except:
print('failed to plot kfmf')
if moments_aux is not None:
try:
self.plot_men()
except:
print('failed to plot men')
try:
self.plot_kfmf_ref()
except:
print('failed to plot ref')
"""
Created on Tue May 26 14:25:40 2020
@author: egorkozlov
"""
import numpy as np
from tiktak import filer
V_base = filer('v_save_base.pkl',0,0)
V_comp = filer('v_save_counterfactual.pkl',0,0)
from setup import ModelSetup
import matplotlib.pyplot as plt
d = {'high education':True}
su_def = ModelSetup(**d)
agrid = su_def.agrid_s
t = 4
a0 = 0.0
ia0 = 0#np.searchsorted(agrid,a0)
sname = 'Female, single'
oname = 'V'
zfgrid = su_def.exogrid.zf_t[t]
zftrend = su_def.pars['f_wage_trend'][t]
names = ['extra assets in base to reach 0 in compare',
'extra assets in compare to reach 0 in base']
def v_compare(v_base, v_compare, he, plot=True, verbose=False):
ia0 = 0
t = 0
a0 = 0.0
from setup import ModelSetup
d = {'high education': he}
su_def = ModelSetup(**d)
agrid = su_def.agrid_s
zfgrid = su_def.exogrid.zf_t[t]
zftrend = su_def.pars['f_wage_trend'][t]
names = [
'extra assets in base to reach 0 in compare',
'extra assets in compare to reach 0 in base'
]
i = 0
wall = list()
aall = list()
for (v0, v1), name in zip([(v_base, v_compare), (v_compare, v_base)],
names):
if verbose:
print('')
print(name)
print('')
print('v0 is larger in {} cases'.format(np.sum(v0 > v1)))
print('v0 is smaller in {} cases'.format(np.sum(v0 < v1)))
wlist = list()
alist = list()
for iz in range(v0.shape[1]):
if verbose: print('iz is {}'.format(iz))
if verbose:
print('wage is {}'.format(np.exp(zfgrid[iz] + zftrend)))
usd_mult = 35
ia_find = np.searchsorted(v0[:, iz], v1[ia0, iz])
if verbose:
print('v1 is {}, v0 is {}'.format(v1[ia0, iz], v0[ia_find,
iz]))
wg = usd_mult * np.exp(zfgrid[iz] + zftrend)
wlist.append(wg)
if verbose: print('the USD wage is {}'.format(wg))
a = usd_mult * (agrid[ia_find] - agrid[ia0])
alist.append(a)
if verbose:
print('the USD equivalent level at the point is {}'.format(a))
if plot:
fig, ax = plt.subplots()
plt.plot(wlist, alist, 'o--k')
plt.title(
'Welfare costs of unplanned pregnancy, \n single women at {}, median income = {:3.1f}'
.format(21 + t, wlist[3]))
plt.xlabel('yearly income, $1000 (2015)')
plt.ylabel('assets equivalent (EV), $1000 (2015)')
ax.grid(True)
#plt.savefig('assets_eq{}.pdf'.format(i))
i += 1
wall.append(np.array(wlist))
aall.append(np.array(alist))
return wall, aall
print(111,psepu,1111,pdivu,11111,dimcu,111111,dimmu)
print(222,psepb,2222,pdivb,22222,dimcb,222222,dimmb)
return fig,fig2,fig3
if __name__ == '__main__':
# some_file.py
import sys
# insert at 1, 0 is the script path (or '' in REPL)
sys.path.insert(1, 'C:/Users\Fabio/Dropbox/cohabitation_marriage')
from setup import ModelSetup
setup = ModelSetup(nogrid=False,sig_zf=0.43,sig_zm=0.43,sigma_psi=0.27,sig_zf_0=0.43,sig_zm_0=0.43)
#v, vf, vm, c, s = v_couple_t0_postmar(setup,2.0,0.0,0.0,+0.1,0.4,div_costs)
am = 1.0 #1.0
af = 0.00 #2.0
zmi = 1 #1.5
zfi = 1 #0.0
psii = 7 #0.0-->0.1
#Creat Graphs of the value of cohab+marriage (unilateral+bilateral) given psi
af0 + am0,
zf0,
zm0,
psi0,
theta,
div_costs,
assets_share_f=ashare_f)
if verbose: print('agreed')
return V, VF, VM, theta
if __name__ == '__main__':
from setup import ModelSetup
setup = ModelSetup(nogrid=True, sig_zf=1.2, sig_zm=1.2, sigma_psi=0.35)
div_costs_mar = DivorceCosts(unilateral_divorce=True,
u_lost_m=0.5,
u_lost_f=0.5)
div_costs_coh = DivorceCosts(unilateral_divorce=True,
u_lost_m=0.0,
u_lost_f=0.0)
#v, vf, vm, c, s = v_couple_t0_postmar(setup,2.0,0.0,0.0,+0.1,0.4,div_costs)
am = 0.0 #1.0
af = 0.0 #2.0
zm = 0.5 #1.5
zf = 0.5 #0.0
psi = 0.1 #0.0-->0.1
class FitPlots(object):
def __init__(self,
*,
base,
targ_mode,
compare=None,
setup=None,
base_name='Model',
compare_name='Data',
graphs_title_add=None,
moments_aux=None,
noplot=False):
if setup is None: self.setup = ModelSetup()
if type(base) is str:
self.moments = filer(base, 0, 0, repeat=False)
else:
self.moments = base
self.targ_mode = targ_mode
self.base_name = base_name
self.compare_name = compare_name
if graphs_title_add:
self.graphs_title_add = '\n ' + graphs_title_add
else:
self.graphs_title_add = ''
if type(compare) is str:
targ_load = filer(compare, 0, 0, repeat=False)
# emulating targets
self.targets = {key: (targ_load[key], 0.0) for key in targ_load}
else:
self.targets = all_targets(targ_mode)
if not noplot:
try:
self.print_things()
except:
print('failed to print')
self.moments_aux = moments_aux
if not noplot: self.plot()
def plot(self):
'''
try:
self.plot_estimates()
except:
print('failed to plot estimates')
try:
self.plot_hazards()
except:
print('failed to plot hazards')
'''
self.plot_trends()
try:
self.plot_cumulative()
except:
print('failed to plot cumulative')
try:
self.plot_by_years_after_marriage()
except:
print('failed to plot by years after marriage')
try:
self.plot_kfmf()
except:
print('failed to plot kfmf')
if self.moments_aux is not None:
try:
self.plot_men()
except:
print('failed to plot men')
try:
self.plot_single_moms()
except:
print('failed to plot single moms')
try:
self.plot_single_moms_new()
except:
print('failed to plot single moms (new)')
try:
self.plot_welfare()
except:
print('failed to plot welfare')
try:
self.plot_kfmf_ref()
except:
print('failed to plot ref')
try:
self.plot_child_welfare()
except:
print('failed to plot child welfare')
plt.show()
def print_things(self):
labels = [
'divorced if k then m and one marriage',
'divorced by years after marriage if kids first, 10',
'divorced if m then k and one marriage',
'divorced by years after marriage if marriage first, 10'
]
for l in labels:
print('{}: base {}, compare {}'.format(l, self.moments[l],
self.targets[l][0]))
def plot_trends(self):
setup = self.setup
tval = np.arange(21, 45)
trend_f = np.zeros_like(tval, dtype=np.float32)
trend_m = np.zeros_like(tval, dtype=np.float32)
for i, t in enumerate(tval):
trend_f[i] = np.exp(setup.pars['f_wage_trend'][i])
trend_m[i] = np.exp(
setup.pars['m_wage_trend'][i - 2]) if i >= 2 else None
fig, ax = plt.subplots()
ax.plot(tval, trend_f, '--b', label='female')
ax.plot(tval, trend_m, '-k', label='male')
ax.set_xlabel('age')
ax.set_ylabel('income (USD 1000s)')
ax.set_title('estimated earnings trends:\n{}'.format(self.base_name))
ax.legend()
ax.grid(True)
plt.savefig('earnings_trend.pdf')
def plot_estimates(self):
setup = self.setup
tval = np.arange(21, 35)
pmeet = np.zeros_like(tval, dtype=np.float64)
ppreg = np.zeros_like(tval, dtype=np.float64)
for i, t in enumerate(tval):
pmeet[i] = setup.pars['pmeet_t'][i]
ppreg[i] = setup.upp_precomputed[i][0]
plt.figure()
#plt.plot(tval,pmeet,label='meeting a partner')
plt.plot(tval, ppreg * pmeet * 100, label='meet and become pregnant')
plt.legend()
plt.title('estimated probabilities' + self.graphs_title_add)
plt.xlabel('age')
plt.ylabel('probability (%)')
def plot_hazards(self, ci=False):
moments = self.moments
targets = self.targets
setup = self.setup
# graph 1: hazard of any marriage
tval = np.arange(23, 36)
names = ['hazard of marriage', 'hazard of new child']
captions = [
"hazard of marriage:\n (% new marriages at [age]) / (% single at [age-1])"
+ self.graphs_title_add,
"hazard of new child:\n (% new births at [age]) / (% childless at [age-1])"
+ self.graphs_title_add
]
for name, cap in zip(names, captions):
haz_model = np.zeros_like(tval, dtype=np.float64)
haz_data = np.zeros_like(tval, dtype=np.float64)
haz_data_lci = np.zeros_like(tval, dtype=np.float64)
haz_data_uci = np.zeros_like(tval, dtype=np.float64)
aux = np.zeros_like(tval, dtype=np.float64)
name_aux = None
for i, t in enumerate(tval):
haz_model[i] = moments['{} at {}'.format(name, t)]
haz_data[i] = targets['{} at {}'.format(name, t)][0]
haz_data_lci[i] = haz_data[i] - 1.96 * targets[
'{} at {}'.format(name, t)][1]
haz_data_uci[i] = haz_data[i] + 1.96 * targets[
'{} at {}'.format(name, t)][1]
if name == 'hazard of marriage':
aux[i] = setup.pars['pmeet_t'][i + 1]
name_aux = 'meeting probability'
elif name == 'hazard of marriage & having a child':
aux[i] = setup.pars['pmeet_t'][
i + 1] * setup.upp_precomputed[i + 1][0]
name_aux = 'meeting + pregnancy probability'
fig, ax = plt.subplots()
plt.plot(tval, haz_model * 100, 'o-b', label=self.base_name)
plt.plot(tval, haz_data * 100, 'o-k', label=self.compare_name)
if ci: plt.plot(tval, haz_data_lci * 100, label='lower 95%')
if ci: plt.plot(tval, haz_data_uci * 100, label='upper 95%')
#if name_aux is not None: plt.plot(tval,aux,label=name_aux)
ax.grid(True)
xticks = [i for i in range(22, 36)]
ax.set_xticks(xticks)
plt.legend()
plt.title(cap)
plt.xlabel('age')
plt.ylabel('hazard (%)')
plt.savefig('{}.pdf'.format(name))
def plot_cumulative(self):
# graph 1: hazard of any marriage
moments, targets, setup = self.moments, self.targets, self.setup
import matplotlib.gridspec as gridspec
tval = np.arange(23, 36)
names = [
'k then m in population', 'm then k in population',
'k then m in sample', 'ever married'
]
captions = [
"Kids First", "Marriage First",
"Relative % of Kids First by age:\n (Kids First) / (Kids First + Marriage First)"
+ self.graphs_title_add, "ever married" + self.graphs_title_add
]
probs_model = []
probs_data = []
for name, cap in zip(names, captions):
prob_model = np.zeros_like(tval, dtype=np.float64)
prob_data = np.zeros_like(tval, dtype=np.float64)
for i, t in enumerate(tval):
prob_model[i] = moments['{} at {}'.format(name, t)]
prob_data[i] = targets['{} at {}'.format(name, t)][0]
probs_model = probs_model + [prob_model * 100]
probs_data = probs_data + [prob_data * 100]
fig = plt.figure() #tight_layout=True)
gs = gridspec.GridSpec(1, 1)
ch = ['o', 'x']
nm = ['KF', 'MF']
for i in range(2):
#if i < 2:
ax = fig.add_subplot(gs[0, 0])
#else:
# ax = fig.add_subplot(gs[:,1])
ax.plot(tval,
probs_model[i],
'{}-b'.format(ch[i]),
label='{}: {}'.format(nm[i], self.base_name.lower()))
ax.plot(tval,
probs_data[i],
'{}-k'.format(ch[i]),
label='{}: {}'.format(nm[i], self.compare_name.lower()))
ax.set_xlabel('age')
ax.set_ylabel('share (%)')
ax.set_title(
'% in female population by age:\nKids First and Marriage First' +
self.graphs_title_add)
ax.legend()
#yticks = [i*5 for i in range(12)]
#ax.set_yticks(yticks)
xticks = [i for i in range(22, 36)]
ax.set_xticks(xticks)
ax.grid(True)
plt.ylim(bottom=-1.0)
plt.savefig('popshares.pdf')
fig, ax = plt.subplots()
ax.plot(tval, probs_model[2], 'o-b', label=self.base_name)
ax.plot(tval, probs_data[2], 'o-k', label=self.compare_name)
ax.set_xlabel('age')
ax.set_ylabel('share (%)')
ax.set_title(captions[2])
ax.legend()
ax.set_xticks(xticks)
ax.grid(True)
plt.ylim(bottom=-1.0)
plt.savefig('relshares.pdf')
fig, ax = plt.subplots()
ax.plot(tval, probs_model[3], 'o-b', label=self.base_name)
ax.plot(tval, probs_data[3], 'o-k', label=self.compare_name)
ax.set_xlabel('age')
ax.set_ylabel('share (%)')
ax.set_title(captions[3])
ax.legend()
ax.set_xticks(xticks)
ax.grid(True)
plt.savefig('evermar.pdf')
def plot_by_years_after_marriage(self):
# graph 1: hazard of any marriage
moments, targets = self.moments, self.targets
yval = np.arange(1, 11)
names = [
'ever kids by years after marriage',
'divorced by years after marriage'
]
captions = [
'% with kids by years after marriage\n (if married)' +
self.graphs_title_add,
'% divorced by years after marriage \n (excluding remarried)' +
self.graphs_title_add
]
fnames = ['kids_by_duration', 'divorced_by_duration']
for name, cap, fname in zip(names, captions, fnames):
prob_model = np.zeros_like(yval, dtype=np.float64)
prob_data = np.zeros_like(yval, dtype=np.float64)
for i, t in enumerate(yval):
try:
prob_model[i] = moments['{}, {}'.format(name, t)]
except:
prob_model[i] = None
print('{}, {} not found in moments'.format(name, t))
try:
prob_data[i] = targets['{}, {}'.format(name, t)][0]
except:
prob_data[i] = None
print('{}, {} not found in targets'.format(name, t))
fig, ax = plt.subplots()
i_data = ~np.isnan(prob_data)
i_model = ~np.isnan(prob_model)
plt.plot(yval[i_model],
prob_model[i_model] * 100,
'o-b',
label=self.base_name)
plt.plot(yval[i_data],
prob_data[i_data] * 100,
'd-k',
label=self.compare_name)
if name == 'ever kids by years after marriage' and self.compare_name == 'data':
plt.plot(yval[i_data][3:],
prob_data[i_data][3:] * 100,
'+',
label='targeted',
linewidth=6,
markersize=14)
#if name_aux is not None: plt.plot(tval,aux,label=name_aux)
plt.legend()
plt.title(cap)
plt.xlabel('years after marriage')
plt.ylabel('share (%)')
ax.grid(True)
ax.set_xticks(yval)
plt.ylim(bottom=-1.0)
plt.savefig('{}.pdf'.format(fname))
def plot_kfmf(self):
# graph 1: hazard of any marriage
moments, targets, setup = self.moments, self.targets, self.setup
yval = np.arange(1, 11)
names = [
'divorced by years after marriage if kids first',
'divorced by years after marriage if marriage first'
]
captions = ['KF', 'MF']
mrkrs = ['x', 'o']
fig, ax = plt.subplots()
for name, cap, mrkr in zip(names, captions, mrkrs):
prob_model = np.zeros_like(yval, dtype=np.float64)
prob_data = np.zeros_like(yval, dtype=np.float64)
for i, t in enumerate(yval):
try:
prob_model[i] = moments['{}, {}'.format(name, t)]
except:
prob_model[i] = None
print('{}, {} not found in moments'.format(name, t))
try:
prob_data[i] = targets['{}, {}'.format(name, t)][0]
except:
prob_data[i] = None
print('{}, {} not found in targets'.format(name, t))
print((cap, prob_model))
i_data = ~np.isnan(prob_data)
i_model = ~np.isnan(prob_model)
plt.plot(yval[i_model],
prob_model[i_model] * 100,
'{}-b'.format(mrkr),
label='{}: {}'.format(cap, self.base_name.lower()))
plt.plot(yval[i_data],
prob_data[i_data] * 100,
'{}-k'.format(mrkr),
label='{}: {}'.format(cap, self.compare_name.lower()))
#if name_aux is not None: plt.plot(tval,aux,label=name_aux)
plt.legend()
plt.title('Divorced by years after marriage by groups' +
self.graphs_title_add)
plt.xlabel('years after marriage')
plt.ylabel('share (%)')
ax.grid(True)
ax.set_xticks(yval)
plt.savefig('div_kfmf.pdf')
# ref_kf = [0.08114558, 0.10186335, 0.11817027, 0.11984021, 0.11823362, 0.11746032, 0.13162393, 0.15057915, 0.14123007, 0.14550265]
# ref_mf = [0. , 0.01083856, 0.02031424, 0.02864134, 0.03768433, 0.04414536, 0.05390435, 0.05976757, 0.0625543 , 0.06301748]
def plot_kfmf_ref(self):
assert False
moments, targets, setup, targ_mode = self.moments, self.targets, self.setup, self.targ_mode
yval = np.arange(1, 11)
#ref_kf = np.array([0.08114558, 0.10186335, 0.11817027, 0.11984021, 0.11823362, 0.11746032, 0.13162393, 0.15057915, 0.14123007, 0.14550265])
#ref_mf = np.array([0. , 0.01083856, 0.02031424, 0.02864134, 0.03768433, 0.04414536, 0.05390435, 0.05976757, 0.0625543 , 0.06301748])
if targ_mode == 'high education':
ref_kf = np.array([
0.03697876, 0.07397737, 0.09510086, 0.12460401, 0.1457346,
0.16467463, 0.18292683, 0.18918919, 0.19642857, 0.21967963
])
ref_mf = np.array([
0., 0.00997299, 0.01585933, 0.02122016, 0.0277903, 0.03765505,
0.04619769, 0.05458448, 0.06238917, 0.06841011
])
else:
ref_kf = np.array([
0.02801336, 0.05332585, 0.08019741, 0.10971467, 0.13539367,
0.1536, 0.17518248, 0.19001832, 0.20896249, 0.23069054
])
ref_mf = np.array([
0., 0.01145147, 0.0232318, 0.0351265, 0.04966411, 0.06725495,
0.08349328, 0.10104955, 0.11824745, 0.13481709
])
here_kf = np.zeros_like(yval, dtype=np.float64)
here_mf = np.zeros_like(yval, dtype=np.float64)
names = [
'divorced by years after marriage if kids first',
'divorced by years after marriage if marriage first'
]
conts = [here_kf, here_mf]
for name, cont in zip(names, conts):
for i, t in enumerate(yval):
try:
cont[i] = moments['{}, {}'.format(name, t)]
except:
cont[i] = None
print('{}, {} not found in moments'.format(name, t))
fig, ax = plt.subplots()
lbl_cmp = '$\phi_s = 0$'
#lbl_cmp = 'equal pay'
plt.plot(yval,
ref_kf * 100,
'x-b',
label='KF: {}'.format(self.base_name.lower()))
plt.plot(yval, here_kf * 100, 'o-b', label='KF: {}'.format(lbl_cmp))
plt.plot(yval,
ref_mf * 100,
'x-k',
label='MF: {}'.format(self.base_name.lower()))
plt.plot(yval, here_mf * 100, 'o--b', label='MF: {}'.format(lbl_cmp))
plt.legend()
plt.title('Divorced by years after marriage: social stigma matters')
#plt.title('Divorced by years after marriage: \n removing gender pay gap')
plt.xlabel('years after marriage')
plt.ylabel('share (%)')
ax.grid(True)
ax.set_xticks(yval)
plt.savefig('div_kfmf_ref.pdf')
def plot_single_moms(self):
# graph 1: hazard of any marriage
moments, targets, setup = self.moments, self.targets, self.setup
tval = np.arange(23, 36)
nm_sm_model = np.zeros_like(tval, dtype=np.float64)
div_sm_model = np.zeros_like(tval, dtype=np.float64)
all_sm_model = np.zeros_like(tval, dtype=np.float64)
nm_sm_data = np.zeros_like(tval, dtype=np.float64)
div_sm_data = np.zeros_like(tval, dtype=np.float64)
all_sm_data = np.zeros_like(tval, dtype=np.float64)
div_childless_data = np.zeros_like(tval, dtype=np.float64)
div_childless_model = np.zeros_like(tval, dtype=np.float64)
nm_childless_data = np.zeros_like(tval, dtype=np.float64)
nm_childless_model = np.zeros_like(tval, dtype=np.float64)
for i, t in enumerate(tval):
nm_sm_model[i] = moments[
'never married and kids in population at {}'.format(t)]
div_sm_model[i] = moments[
'divorced and kids in population at {}'.format(t)]
all_sm_model[i] = nm_sm_model[i] + div_sm_model[i]
nm_sm_data[i] = targets[
'never married and kids in population at {}'.format(t)][0]
div_sm_data[i] = targets[
'divorced and kids in population at {}'.format(t)][0]
all_sm_data[i] = nm_sm_data[i] + div_sm_data[i]
nm_childless_data[i] = targets[
'never married and no kids in population at {}'.format(t)][0]
nm_childless_model[i] = moments[
'never married and no kids in population at {}'.format(t)]
div_childless_data[i] = targets[
'divorced and no kids in population at {}'.format(t)][0]
div_childless_model[i] = moments[
'divorced and no kids in population at {}'.format(t)]
fig, ax = plt.subplots()
ax.plot(tval, 100 * all_sm_model, 'o-b', label=self.base_name)
ax.plot(tval, 100 * all_sm_data, 'o-k', label=self.compare_name)
ax.set_xlabel('age')
ax.set_ylabel('share (%)')
ax.set_title('Single mothers in population')
ax.legend()
ax.set_xticks(tval)
plt.ylim(bottom=-1.0)
ax.grid(True)
plt.savefig('fit_single_mothers_all.pdf')
fig, ax = plt.subplots()
ax.plot(tval, 100 * div_sm_model, 'o-b', label=self.base_name)
ax.plot(tval, 100 * div_sm_data, 'd-k', label=self.compare_name)
if self.compare_name == 'data':
ax.plot(tval[5:],
100 * div_sm_data[5:],
'+k',
label='targeted',
linewidth=6,
markersize=14)
ax.set_xlabel('age')
ax.set_ylabel('share (%)')
ax.set_title('Divorced with kids in population')
ax.legend()
ax.set_xticks(tval)
plt.ylim(bottom=-0.2)
ax.grid(True)
plt.savefig('fit_single_mothers_div.pdf')
fig, ax = plt.subplots()
ax.plot(tval, 100 * nm_sm_model, 'o-b', label=self.base_name)
ax.plot(tval, 100 * nm_sm_data, 'd-k', label=self.compare_name)
if self.compare_name == 'data':
ax.plot(tval[5:],
100 * nm_sm_data[5:],
'+k',
label='targeted',
linewidth=6,
markersize=14)
ax.set_xlabel('age')
ax.set_ylabel('share (%)')
plt.ylim(bottom=-0.2)
ax.set_title('Never married with kids in population')
ax.legend()
ax.set_xticks(tval)
ax.grid(True)
plt.savefig('fit_single_mothers_nm.pdf')
fig, ax = plt.subplots()
ax.plot(tval, 100 * div_childless_model, 'o-b', label=self.base_name)
ax.plot(tval, 100 * div_childless_data, 'd-k', label=self.compare_name)
if self.compare_name == 'data':
ax.plot(tval[5:],
100 * div_childless_data[5:],
'+k',
label='targeted',
linewidth=6,
markersize=14)
ax.set_xlabel('age')
ax.set_ylabel('share (%)')
ax.set_title('Divorced witout kids in population')
ax.legend()
ax.set_xticks(tval)
plt.ylim(bottom=-0.2)
ax.grid(True)
plt.savefig('fit_single_nonmothers_div.pdf')
fig, ax = plt.subplots()
ax.plot(tval, 100 * nm_childless_model, 'o-b', label=self.base_name)
ax.plot(tval, 100 * nm_childless_data, 'd-k', label=self.compare_name)
if self.compare_name == 'data':
ax.plot(tval[0:],
100 * nm_childless_data[0:],
'+k',
label='targeted',
linewidth=6,
markersize=14)
ax.set_xlabel('age')
ax.set_ylabel('share (%)')
ax.set_title('Never married without kids in population')
ax.legend()
ax.set_xticks(tval)
plt.ylim(bottom=-1.0)
ax.grid(True)
plt.savefig('fit_single_nonmothers_nm.pdf')
def plot_single_moms_new(self):
# graph 1: hazard of any marriage
moments, targets, setup = self.moments, self.targets, self.setup
tval = np.arange(25, 36)
all_sm_model = np.zeros_like(tval, dtype=np.float64)
rel_sm_model = np.zeros_like(tval, dtype=np.float64)
abortions_model = np.zeros_like(tval, dtype=np.float64)
all_sm_data = np.zeros_like(tval, dtype=np.float64)
rel_sm_data = np.zeros_like(tval, dtype=np.float64)
abortions_data = np.zeros_like(tval, dtype=np.float64)
for i, t in enumerate(tval):
all_sm_model[i] = moments['single mothers in total at {}'.format(
t)]
all_sm_data[i] = targets['single mothers in total at {}'.format(
t)][0]
rel_sm_model[i] = moments[
'single mothers among mothers at {}'.format(t)]
rel_sm_data[i] = targets[
'single mothers among mothers at {}'.format(t)][0]
all_sm_model[i] = moments['single mothers in total at {}'.format(
t)]
all_sm_data[i] = targets['single mothers in total at {}'.format(
t)][0]
abortions_model[i] = moments[
'aborted unplanned pregnancies at {}'.format(t)]
abortions_data[i] = targets[
'aborted unplanned pregnancies at {}'.format(t)][0]
fig, ax = plt.subplots()
ax.plot(tval, 100 * all_sm_model, 'o-b', label=self.base_name)
ax.plot(tval, 100 * all_sm_data, 'o-k', label=self.compare_name)
ax.set_xlabel('age')
ax.set_ylabel('share (%)')
ax.set_title('Single mothers in population')
ax.legend()
ax.set_xticks(tval)
ax.grid(True)
am = 100 * moments['unplanned pregnancies aborted']
ad = 100 * targets['unplanned pregnancies aborted'][0]
fig, ax = plt.subplots()
ax.plot(tval,
100 * rel_sm_model,
'o-b',
label=self.base_name + ', abortions {:02.3g}%'.format(am))
ax.plot(tval,
100 * rel_sm_data,
'o-k',
label=self.compare_name + ', abortions {:02.3g}%'.format(ad))
ax.set_xlabel('age')
ax.set_ylabel('share (%)')
ax.set_title('Single mothers among mothers + abortions')
ax.legend()
ax.set_xticks(tval)
ax.grid(True)
plt.savefig('sm_among_mothers.pdf')
fig, ax = plt.subplots()
ax.plot(tval, abortions_model, 'o-b', label=self.base_name)
ax.plot(tval, abortions_data, 'o-k', label=self.compare_name)
ax.set_xlabel('age')
ax.set_ylabel('share (%)')
ax.set_title('Share of abortions')
ax.legend()
ax.set_xticks(tval)
ax.grid(True)
def plot_welfare(self, a_mult=1, noplot=False):
# singles
z_fem = self.setup.exogrid.zf_t[0]
w_fem = np.exp(z_fem + self.setup.pars['f_wage_trend'][0])
z_mal = self.setup.exogrid.zm_t[0]
w_mal = np.exp(z_mal + self.setup.pars['f_wage_trend'][0])
moments, targets = self.moments, self.targets
# males and females
a_female = \
v_compare(self.setup.agrid_s,
moments['value function: female, single, all assets'],
targets['value function: female, single, all assets'][0],
a_mult=a_mult)
a_male = \
v_compare(self.setup.agrid_s,
moments['value function: male, single, all assets'],
targets['value function: male, single, all assets'][0],
a_mult=a_mult)
if not noplot:
print(a_male)
print(a_female)
fig, ax = plt.subplots()
plt.plot(z_fem,
a_female,
'o-b',
label='{} - {}'.format(self.base_name, self.compare_name))
ax.grid(True)
xticks = z_fem
ax.set_xticks(xticks)
plt.legend()
plt.title('Welfare differences: female, 21, no kids, no assets')
plt.xlabel('female productivity')
plt.ylabel('asset equivalent variation (2016 USD, 1000s)')
plt.savefig('welfare_female.pdf')
fig, ax = plt.subplots()
plt.plot(z_mal,
a_male,
'o-b',
label='{} - {}'.format(self.base_name, self.compare_name))
ax.grid(True)
xticks = z_mal
ax.set_xticks(xticks)
plt.legend()
plt.title('Welfare differences: male, 23, no kids, no assets')
plt.xlabel('male productivity')
plt.ylabel('asset equivalent variation (2016 USD, 1000s)')
plt.savefig('welfare_male.pdf')
# couples
# transofrm the state
ipsi = np.arange(self.setup.pars['n_psi'])
imal = 2 * np.ones_like(ipsi)
ifem = 3 * np.ones_like(ipsi)
iexo, _, _, _ = self.setup.all_indices(0, (ifem, imal, ipsi))
a_couple = \
v_compare(self.setup.agrid_c,
moments['value function: couple, no children, all assets'][:,iexo],
targets['value function: couple, no children, all assets'][0][:,iexo],
a_mult=a_mult)
if not noplot:
psi = self.setup.exogrid.psi_t[0]
fig, ax = plt.subplots()
#plt.plot(psi,v_couple_base_val[iexo],'o-b',label=self.base_name)
#if v_couple_compare_val is not None: plt.plot(psi,v_couple_compare_val[iexo],'o-k',label=self.compare_name)
plt.plot(psi,
a_couple,
'o-b',
label='{} - {}'.format(self.base_name, self.compare_name))
ax.grid(True)
xticks = np.linspace(psi.min(), psi.max(), 7)
ax.set_xticks(xticks)
plt.legend()
plt.title('Welfare comparison: couple, 21/23, no kids, no assets')
plt.xlabel('love shock')
plt.ylabel('asset equivalent variation (2016 USD, 1000s)')
if noplot: return a_female, a_male, a_couple
def plot_child_welfare(self):
# singles
aval = np.arange(25, 36)
base_val = np.array([
self.moments['child welfare if born at {}, median'.format(a)]
for a in aval
])
compare_val = np.array([
self.targets['child welfare if born at {}, median'.format(a)][0]
for a in aval
])
change_mean = 100 * (
self.moments['welfare of all children born, mean'] -
self.targets['welfare of all children born, mean'][0]
) / self.moments['welfare of all children born, mean']
change_median = 100 * (
self.moments['welfare of all children born, median'] -
self.targets['welfare of all children born, median'][0]
) / self.moments['welfare of all children born, median']
fig, ax = plt.subplots()
plt.plot(aval,
base_val - compare_val,
'o-b',
label='{} - {}'.format(self.base_name, self.compare_name))
ax.grid(True)
xticks = aval
ax.set_xticks(xticks)
plt.legend()
plt.title(
"Child consumption equivalent differences, by mother's are\n changes mean = {:02.1f}%, median = {:02.1f}%"
.format(change_mean, change_median))
plt.xlabel('age the child is born')
plt.ylabel('child consumption equivalent')
def plot_men(self):
moments, targets = self.moments, self.targets
tval = np.arange(24, 36)
names = [
'men, relative income just married / single',
'men, relative income with kids / no kids'
]
captions = [
'men, relative income just married / single',
'married men, relative income with kids / no kids'
]
fnames = ['men_mar_ratio', 'men_kids_ratio']
for name, cap, fname in zip(names, captions, fnames):
haz_model = np.zeros_like(tval, dtype=np.float64)
haz_data = np.zeros_like(tval, dtype=np.float64)
haz_data_lci = np.zeros_like(tval, dtype=np.float64)
haz_data_uci = np.zeros_like(tval, dtype=np.float64)
for i, t in enumerate(tval):
haz_model[i] = moments['{} at {}'.format(name, t)]
haz_data[i] = targets['{} at {}'.format(name, t)][0]
haz_data_lci[i] = haz_data[i] - 1.96 * targets[
'{} at {}'.format(name, t)][1]
haz_data_uci[i] = haz_data[i] + 1.96 * targets[
'{} at {}'.format(name, t)][1]
fig, ax = plt.subplots()
plt.plot(tval, haz_model * 100, 'o-b', label=self.base_name)
plt.plot(tval, haz_data * 100, 'o-k', label=self.compare_name)
ax.grid(True)
xticks = tval #[i for i in range(24,36)]
ax.set_xticks(xticks)
plt.legend()
plt.title(cap)
plt.xlabel('age')
plt.ylabel('ratio (%)')
plt.savefig('{}.pdf'.format(fname))