Skip to content

Commit

Permalink
Consolidated UBI>0 tests in test_net_taxes() in test_tax.py
Browse files Browse the repository at this point in the history
  • Loading branch information
@rickecon
rickecon committed May 25, 2021
1 parent e17695d commit 047dd74
Showing 1 changed file with 111 additions and 128 deletions.
239 changes: 111 additions & 128 deletions ogusa/tests/test_tax.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -512,7 +512,7 @@ def test_get_biz_tax():

'''
-------------------------------------------------------------------------------
Test baseline tax.py net_taxes() function
Test tax.py net_taxes() function
-------------------------------------------------------------------------------
'''
# Set parameter class for each case
Expand Down Expand Up @@ -671,6 +671,98 @@ def test_get_biz_tax():
ubi3 = np.zeros((p3.T, p3.S, p3.J))
ubi4 = np.zeros((p4.T, p4.S, p4.J))
ubi5 = np.zeros((p5.T, p5.S, p5.J))

p_u = Specifications()
new_param_values_ubi = {
'T': 3,
'S': 3,
'J': 2,
'lambdas': [0.65, 0.35],
'eta': (np.ones((3, 2)) / (3 * 2)),
'ubi_nom_017': 1000,
'ubi_nom_1864': 1500,
'ubi_nom_65p': 500
}
p_u.update_specifications(new_param_values_ubi)
p_u.tax_func_type = 'DEP'
p_u.e = np.array([[0.3, 0.2], [0.5, 0.4], [0.45, 0.3]])
p_u.h_wealth = np.ones(p_u.T + p_u.S) * 1
p_u.p_wealth = np.ones(p_u.T + p_u.S) * 2
p_u.m_wealth = np.ones(p_u.T + p_u.S) * 3
p_u.tau_payroll = np.ones(p_u.T + p_u.S) * 0.15
p_u.tau_bq = np.ones(p_u.T + p_u.S) * 0.1
p_u.retire = (np.ones(p_u.T + p_u.S) * 2).astype(int)

factor_u = 105000
ubi_u = p_u.ubi_nom_array / factor_u

# set variables and other parameters for each case
r9 = 0.04
w9 = 1.2
b9 = np.array([0.4, 0.3, 0.5])
n9 = np.array([0.8, 0.4, 0.7])
BQ9 = np.array([0.3])
bq9 = BQ9 / p_u.lambdas[0]
tr9 = np.array([0.12])
theta9 = np.array([0.225])
etr_params9_old = np.reshape(np.array([
[0.001, 0.002, 0.003, 0.0015, 0.8, -0.14, 0.8, -0.15, 0.15,
0.16, -0.15, 0.83],
[0.001, 0.002, 0.003, 0.0015, 0.8, -0.14, 0.8, -0.15, 0.15,
0.16, -0.15, 0.83],
[0.001, 0.002, 0.003, 0.0015, 0.8, -0.14, 0.8, -0.15, 0.15,
0.16, -0.15, 0.83]]), (1, p_u.S, 12))
etr_params9 = etr_params9_old.copy()
etr_params9[:, :, 5] = etr_params9_old[:, :, 6]
etr_params9[:, :, 6] = etr_params9_old[:, :, 11]
etr_params9[:, :, 7] = etr_params9_old[:, :, 5]
etr_params9[:, :, 8] = etr_params9_old[:, :, 7]
etr_params9[:, :, 9] = etr_params9_old[:, :, 8]
etr_params9[:, :, 10] = etr_params9_old[:, :, 9]
etr_params9[:, :, 11] = etr_params9_old[:, :, 10]
j9 = 0
shift9 = True
method9 = 'SS'

r10 = r9
w10 = w9
b10 = b9
n10 = n9
BQ10 = BQ9
bq10 = bq9
tr10 = tr9
theta10 = theta9
etr_params10 = etr_params9
j10 = 0
shift10 = True
method10 = 'TPI_scalar'

r11 = np.array([0.04, 0.045, 0.04])
w11 = np.array([1.2, 1.3, 1.1])
b11 = np.tile(np.reshape(np.array([0.4, 0.3, 0.5]), (1, p_u.S, 1)),
(p_u.T, 1, p_u.J))
n11 = np.tile(np.reshape(np.array([0.8, 0.4, 0.7]), (1, p_u.S, 1)),
(p_u.T, 1, p_u.J))
BQ11 = np.array([0.3, 0.4, 0.45])
bq11 = np.tile(np.reshape(BQ11 / p_u.lambdas[0], (p_u.T, 1)), (1, p_u.S))
tr11 = np.tile(np.reshape(np.array([0.12, 0.1, 0.11]), (p_u.T, 1)),
(1, p_u.S))
theta11 = theta9
etr_params11_old = np.tile(np.reshape(np.array(
[0.001, 0.002, 0.003, 0.0015, 0.8, -0.14, 0.8, -0.15, 0.15, 0.16,
-0.15, 0.83]), (1, 1, 12)), (p_u.T, p_u.S, 1))
etr_params11 = etr_params11_old.copy()
etr_params11[:, :, 5] = etr_params11_old[:, :, 6]
etr_params11[:, :, 6] = etr_params11_old[:, :, 11]
etr_params11[:, :, 7] = etr_params11_old[:, :, 5]
etr_params11[:, :, 8] = etr_params11_old[:, :, 7]
etr_params11[:, :, 9] = etr_params11_old[:, :, 8]
etr_params11[:, :, 10] = etr_params11_old[:, :, 9]
etr_params11[:, :, 11] = etr_params11_old[:, :, 10]
j11 = 0
shift11 = True
method11 = 'TPI'

expected1 = np.array([0.47374766, -0.09027663, 0.03871394])
expected2 = np.array([0.20374766, -0.09027663, 0.03871394])
expected3 = np.array([[0.473747659, -0.090276635, 0.038713941],
Expand Down Expand Up @@ -717,6 +809,11 @@ def test_get_biz_tax():
[[0.31524401 - 0.061538462, 0.21763702 - 0.12857143],
[0.34545346 - 0.061538462, 0.39350691 - 0.12857143],
[0.15958077 - 0.061538462, -0.0482051 - 0.12857143]]])
expected9 = np.array([0.28384671, -0.07461627, 0.15144631])
expected10 = np.array([0.01384671, -0.07461627, 0.15144631])
expected11 = np.array([[0.30869878, -0.0497642, 0.17629838],
[0.39311429, 0.00794409, 0.24575229],
[0.35257605, 0.02042006, 0.23553789]])

test_data = [(r1, w1, b1, n1, bq1, factor, tr1, ubi1[0, :, :], theta1, None,
j1, shift1, method1, p1.e[:, j1], etr_params1[-1, :, :], p1,
Expand All @@ -736,7 +833,16 @@ def test_get_biz_tax():
(r5, w5, b5, n5, bq5, factor, tr5, ubi5, theta5, 0, j5, shift5,
method5, p5.e, etr_params5, p7, expected7),
(r5, w5, b5, n5, bq5, factor, tr5, ubi5, theta5, 0, j5, shift5,
method5, p5.e, etr_params5, p8, expected8)]
method5, p5.e, etr_params5, p8, expected8),
(r9, w9, b9, n9, bq9, factor_u, tr9, ubi_u[0, :, j9], theta9,
None, j9, shift9, method9, p_u.e[:, j9], etr_params9[-1, :, :],
p_u, expected9),
(r10, w10, b10, n10, bq10, factor_u, tr10, ubi_u[0, :, j10],
theta10, None, j10, shift10, method10, p_u.e[:, j10],
etr_params10, p_u, expected10),
(r11, w11, b11[:, :, j11], n11[:, :, j11], bq11, factor_u, tr11,
ubi_u[:p_u.T, :, j11], theta11, 0, j11, shift11, method11,
p_u.e[:, j11], etr_params11, p_u, expected11)]


@pytest.mark.parametrize('r,w,b,n,bq,factor,tr,ubi,theta,t,j,shift,method,'
Expand All @@ -746,135 +852,12 @@ def test_get_biz_tax():
'TPI shift = False', 'TPI 3D',
'TPI 3D,vary tau_bq',
'TPI 3D,vary retire',
'TPI 3D,vary replacement rate'])
'TPI 3D,vary replacement rate',
'SS UBI>0', 'TPI scalar UBI>0',
'TPI UBI>0'])
def test_net_taxes(r, w, b, n, bq, factor, tr, ubi, theta, t, j, shift,
method, e, etr_params, p, expected):
# Test function that computes total net taxes for the household
net_taxes = tax.net_taxes(r, w, b, n, bq, factor, tr, ubi, theta, t,
j, shift, method, e, etr_params, p)
assert np.allclose(net_taxes, expected)


'''
-------------------------------------------------------------------------------
Local UBI > 0 test of net_taxes() function
-------------------------------------------------------------------------------
'''
# Set parameter class for each case
p_u = Specifications()
new_param_values_ubi = {
'T': 3,
'S': 3,
'J': 2,
'lambdas': [0.65, 0.35],
'eta': (np.ones((3, 2)) / (3 * 2)),
'ubi_nom_017': 1000,
'ubi_nom_1864': 1500,
'ubi_nom_65p': 500
}
p_u.update_specifications(new_param_values_ubi)
p_u.tax_func_type = 'DEP'
p_u.e = np.array([[0.3, 0.2], [0.5, 0.4], [0.45, 0.3]])
p_u.h_wealth = np.ones(p_u.T + p_u.S) * 1
p_u.p_wealth = np.ones(p_u.T + p_u.S) * 2
p_u.m_wealth = np.ones(p_u.T + p_u.S) * 3
p_u.tau_payroll = np.ones(p_u.T + p_u.S) * 0.15
p_u.tau_bq = np.ones(p_u.T + p_u.S) * 0.1
p_u.retire = (np.ones(p_u.T + p_u.S) * 2).astype(int)

factor = 105000
ubi_u = p_u.ubi_nom_array / factor

# set variables and other parameters for each case
r1 = 0.04
w1 = 1.2
b1 = np.array([0.4, 0.3, 0.5])
n1 = np.array([0.8, 0.4, 0.7])
BQ1 = np.array([0.3])
bq1 = BQ1 / p_u.lambdas[0]
tr1 = np.array([0.12])
theta1 = np.array([0.225])
etr_params1_old = np.reshape(np.array([
[0.001, 0.002, 0.003, 0.0015, 0.8, -0.14, 0.8, -0.15, 0.15,
0.16, -0.15, 0.83],
[0.001, 0.002, 0.003, 0.0015, 0.8, -0.14, 0.8, -0.15, 0.15,
0.16, -0.15, 0.83],
[0.001, 0.002, 0.003, 0.0015, 0.8, -0.14, 0.8, -0.15, 0.15,
0.16, -0.15, 0.83]]), (1, p_u.S, 12))
etr_params1 = etr_params1_old.copy()
etr_params1[:, :, 5] = etr_params1_old[:, :, 6]
etr_params1[:, :, 6] = etr_params1_old[:, :, 11]
etr_params1[:, :, 7] = etr_params1_old[:, :, 5]
etr_params1[:, :, 8] = etr_params1_old[:, :, 7]
etr_params1[:, :, 9] = etr_params1_old[:, :, 8]
etr_params1[:, :, 10] = etr_params1_old[:, :, 9]
etr_params1[:, :, 11] = etr_params1_old[:, :, 10]
j1 = 0
shift1 = True
method1 = 'SS'

r2 = r1
w2 = w1
b2 = b1
n2 = n1
BQ2 = BQ1
bq2 = bq1
tr2 = tr1
theta2 = theta1
etr_params2 = etr_params1
j2 = 0
shift2 = True
method2 = 'TPI_scalar'

r3 = np.array([0.04, 0.045, 0.04])
w3 = np.array([1.2, 1.3, 1.1])
b3 = np.tile(np.reshape(np.array([0.4, 0.3, 0.5]), (1, p_u.S, 1)),
(p_u.T, 1, p_u.J))
n3 = np.tile(np.reshape(np.array([0.8, 0.4, 0.7]), (1, p_u.S, 1)),
(p_u.T, 1, p_u.J))
BQ3 = np.array([0.3, 0.4, 0.45])
bq3 = np.tile(np.reshape(BQ3 / p_u.lambdas[0], (p_u.T, 1)), (1, p_u.S))
tr3 = np.tile(np.reshape(np.array([0.12, 0.1, 0.11]), (p_u.T, 1)),
(1, p_u.S))
theta3 = theta1
etr_params3_old = np.tile(np.reshape(np.array(
[0.001, 0.002, 0.003, 0.0015, 0.8, -0.14, 0.8, -0.15, 0.15, 0.16,
-0.15, 0.83]), (1, 1, 12)), (p_u.T, p_u.S, 1))
etr_params3 = etr_params3_old.copy()
etr_params3[:, :, 5] = etr_params3_old[:, :, 6]
etr_params3[:, :, 6] = etr_params3_old[:, :, 11]
etr_params3[:, :, 7] = etr_params3_old[:, :, 5]
etr_params3[:, :, 8] = etr_params3_old[:, :, 7]
etr_params3[:, :, 9] = etr_params3_old[:, :, 8]
etr_params3[:, :, 10] = etr_params3_old[:, :, 9]
etr_params3[:, :, 11] = etr_params3_old[:, :, 10]
j3 = 0
shift3 = True
method3 = 'TPI'

expected1 = np.array([0.28384671, -0.07461627, 0.15144631])
expected2 = np.array([0.01384671, -0.07461627, 0.15144631])
expected3 = np.array([[0.30869878, -0.0497642, 0.17629838],
[0.39311429, 0.00794409, 0.24575229],
[0.35257605, 0.02042006, 0.23553789]])

test_data = [(r1, w1, b1, n1, bq1, factor, tr1, ubi_u[0, :, j1], theta1, None,
j1, shift1, method1, p_u.e[:, j1], etr_params1[-1, :, :], p_u,
expected1),
(r2, w2, b2, n2, bq2, factor, tr2, ubi_u[0, :, j2], theta2, None,
j2, shift2, method2, p_u.e[:, j2], etr_params2, p_u, expected2),
(r3, w3, b3[:, :, j3], n3[:, :, j3], bq3, factor, tr3,
ubi_u[:p_u.T, :, j3], theta3, 0, j3, shift3, method3, p_u.e[:, j3],
etr_params3, p_u, expected3)]

@pytest.mark.local
@pytest.mark.parametrize('r,w,b,n,bq,factor,tr,ubi,theta,t,j,shift,method,'
+ 'e,etr_params,p,expected',
test_data, ids=['SS UBI>0', 'TPI scalar UBI>0',
'TPI UBI>0'])
def test_net_taxes_ubi(r, w, b, n, bq, factor, tr, ubi, theta, t, j, shift,
method, e, etr_params, p, expected):
# Test function that computes total net taxes for the household
net_taxes = tax.net_taxes(r, w, b, n, bq, factor, tr, ubi, theta, t,
j, shift, method, e, etr_params, p)
assert np.allclose(net_taxes, expected)

0 comments on commit 047dd74

Please sign in to comment.