This notebook contains an example for how to use the taxbrain python package
# # Install conda, taxbrain, and taxcalc if in Google Colab.
import sys
if 'google.colab' in sys.modules and 'taxbrain' not in sys.modules:
# Install taxbrain and dependencies
!pip install taxbrain &> /dev/null
!pip install taxcalc &> /dev/null
!pip install pypandoc &> /dev/null
!pip install -U pandas &> /dev/null # make sure pandas is up to date
from taxbrain import TaxBrain, differences_plot, distribution_plot
reform_url = "https://raw.githubusercontent.com/PSLmodels/Tax-Calculator/master/taxcalc/reforms/Larson2019.json"
start_year = 2021
end_year = 2030
Static Reform#
After importing the TaxBrain class from the taxbrain package, we initiate an instance of the class by specifying the start and end year of the anlaysis, which microdata to use, and a policy reform. Additional arguments can be used to specify econoimc assumptions and individual behavioral elasticites.
Once the class has been initiated, the run() method will handle executing each model
tb_static = TaxBrain(start_year, end_year, microdata="CPS", reform=reform_url)
tb_static.run()
---------------------------------------------------------------------------
KeyboardInterrupt Traceback (most recent call last)
File /usr/share/miniconda/envs/taxbrain-dev/lib/python3.12/site-packages/numba/core/serialize.py:30, in _numba_unpickle(address, bytedata, hashed)
27 _unpickled_memo = {}
---> 30 def _numba_unpickle(address, bytedata, hashed):
31 """Used by `numba_unpickle` from _helperlib.c
32
33 Parameters
(...)
42 unpickled object
43 """
KeyboardInterrupt:
The above exception was the direct cause of the following exception:
SystemError Traceback (most recent call last)
Cell In[4], line 2
1 tb_static = TaxBrain(start_year, end_year, microdata="CPS", reform=reform_url)
----> 2 tb_static.run()
File ~/work/Tax-Brain/Tax-Brain/taxbrain/taxbrain.py:178, in TaxBrain.run(self, varlist, client, num_workers)
176 if self.verbose:
177 print("Running static simulations")
--> 178 self._static_run(
179 varlist, base_calc, reform_calc, client, num_workers
180 )
181 del base_calc, reform_calc
183 setattr(self, "has_run", True)
File ~/work/Tax-Brain/Tax-Brain/taxbrain/taxbrain.py:440, in TaxBrain._static_run(self, varlist, base_calc, reform_calc, client, num_workers)
434 lazy_values = []
435 for yr in range(self.start_year, self.end_year + 1):
436 lazy_values.extend(
437 [
438 delayed(self._taxcalc_advance(base_calc, varlist, yr)),
439 delayed(
--> 440 self._taxcalc_advance(
441 reform_calc, varlist, yr, reform=True
442 )
443 ),
444 ]
445 )
446 if client:
447 futures = client.compute(lazy_values, num_workers=num_workers)
File ~/work/Tax-Brain/Tax-Brain/taxbrain/taxbrain.py:392, in TaxBrain._taxcalc_advance(self, calc, varlist, year, reform)
384 if reform:
385 calc = dist_corp(
386 calc,
387 self.corp_revenue,
(...)
390 self.ci_params,
391 )
--> 392 calc.calc_all()
393 df = calc.dataframe(varlist)
395 return df
File /usr/share/miniconda/envs/taxbrain-dev/lib/python3.12/site-packages/taxcalc/calculator.py:172, in Calculator.calc_all(self, zero_out_calc_vars)
170 UBI(self.__policy, self.__records)
171 BenefitPrograms(self)
--> 172 self._calc_one_year(zero_out_calc_vars)
173 BenefitSurtax(self)
174 BenefitLimitation(self)
File /usr/share/miniconda/envs/taxbrain-dev/lib/python3.12/site-packages/taxcalc/calculator.py:1435, in Calculator._calc_one_year(self, zero_out_calc_vars)
1433 del item_cvar
1434 # Calculate taxes with optimal itemized deduction
-> 1435 self._taxinc_to_amt()
1436 F2441(self.__policy, self.__records)
1437 EITC(self.__policy, self.__records)
File /usr/share/miniconda/envs/taxbrain-dev/lib/python3.12/site-packages/taxcalc/calculator.py:1365, in Calculator._taxinc_to_amt(self)
1361 """
1362 Call TaxInc through AMT functions.
1363 """
1364 TaxInc(self.__policy, self.__records)
-> 1365 SchXYZTax(self.__policy, self.__records)
1366 GainsTax(self.__policy, self.__records)
1367 AGIsurtax(self.__policy, self.__records)
File /usr/share/miniconda/envs/taxbrain-dev/lib/python3.12/site-packages/taxcalc/decorators.py:325, in iterate_jit.<locals>.make_wrapper.<locals>.wrapper(*args, **kwargs)
322 eval(func_code, # pylint: disable=eval-used
323 {"applied_f": applied_jitted_f}, fakeglobals)
324 high_level_fn = fakeglobals['hl_func']
--> 325 ans = high_level_fn(*args, **kwargs)
326 return ans
File <string>:12, in hl_func(pm, pf)
SystemError: CPUDispatcher(<function ap_func at 0x7f3ebcc29bc0>) returned a result with an exception set
Once the calculators have been run, you can produce a number of tables, including a weighted total of a given variable each year under both current law and the user reform.
print("Combined Tax Liability Over the Budget Window")
tb_static.weighted_totals("combined")
Combined Tax Liability Over the Budget Window
| 2021 | 2022 | 2023 | 2024 | 2025 | 2026 | 2027 | 2028 | 2029 | 2030 | |
|---|---|---|---|---|---|---|---|---|---|---|
| Base | 2.269727e+12 | 3.067767e+12 | 3.246623e+12 | 3.395728e+12 | 3.575938e+12 | 3.979716e+12 | 4.161639e+12 | 4.344151e+12 | 4.543397e+12 | 4.749701e+12 |
| Reform | 2.326818e+12 | 3.145972e+12 | 3.342312e+12 | 3.509844e+12 | 3.710034e+12 | 4.126726e+12 | 4.329940e+12 | 4.534780e+12 | 4.744452e+12 | 4.961574e+12 |
| Difference | 5.709077e+10 | 7.820500e+10 | 9.568834e+10 | 1.141166e+11 | 1.340957e+11 | 1.470099e+11 | 1.683013e+11 | 1.906291e+11 | 2.010544e+11 | 2.118730e+11 |
If you are interested in a detailed look on the reform’s effect, you can produce a differences table for a given year.
print("Differences Table")
tb_static.differences_table(start_year, "weighted_deciles", "combined")
Differences Table
| count | tax_cut | perc_cut | tax_inc | perc_inc | mean | tot_change | share_of_change | ubi | benefit_cost_total | benefit_value_total | pc_aftertaxinc | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0-10n | 0.102049 | 0.000000 | 0.000000 | 0.035165 | 34.458500 | 8.218450 | 0.000839 | 0.001469 | 0.0 | 0.0 | 0.0 | 0.006305 |
| 0-10z | 8.390627 | 0.000000 | 0.000000 | 0.000000 | 0.000000 | 0.000000 | 0.000000 | 0.000000 | 0.0 | 0.0 | 0.0 | 0.000000 |
| 0-10p | 12.214010 | 0.000000 | 0.000000 | 5.510201 | 45.113777 | 3.693062 | 0.045107 | 0.079009 | 0.0 | 0.0 | 0.0 | -0.030271 |
| 10-20 | 20.708304 | 0.000000 | 0.000000 | 15.031974 | 72.589110 | 17.199769 | 0.356178 | 0.623880 | 0.0 | 0.0 | 0.0 | -0.049703 |
| 20-30 | 20.706921 | 0.019010 | 0.091803 | 13.230015 | 63.891753 | 27.008003 | 0.559253 | 0.979585 | 0.0 | 0.0 | 0.0 | -0.047439 |
| 30-40 | 20.707534 | 0.259493 | 1.253135 | 11.850040 | 57.225744 | 29.340367 | 0.607567 | 1.064212 | 0.0 | 0.0 | 0.0 | -0.036801 |
| 40-50 | 20.706181 | 1.020739 | 4.929636 | 12.904516 | 62.322051 | 27.887624 | 0.577446 | 1.011453 | 0.0 | 0.0 | 0.0 | -0.014320 |
| 50-60 | 20.708634 | 2.983512 | 14.407092 | 12.853635 | 62.068967 | -36.541184 | -0.756718 | -1.325465 | 0.0 | 0.0 | 0.0 | 0.110734 |
| 60-70 | 20.707540 | 3.810788 | 18.402900 | 12.691879 | 61.291099 | -155.091891 | -3.211572 | -5.625378 | 0.0 | 0.0 | 0.0 | 0.261506 |
| 70-80 | 20.705701 | 4.042144 | 19.521890 | 13.340665 | 64.429912 | -198.758187 | -4.115428 | -7.208569 | 0.0 | 0.0 | 0.0 | 0.271550 |
| 80-90 | 20.709850 | 4.418265 | 21.334121 | 14.297969 | 69.039464 | -427.711584 | -8.857843 | -15.515367 | 0.0 | 0.0 | 0.0 | 0.416827 |
| 90-100 | 20.708024 | 1.431391 | 6.912254 | 17.352383 | 83.795459 | 3471.405278 | 71.885943 | 125.915170 | 0.0 | 0.0 | 0.0 | -1.292653 |
| ALL | 207.075375 | 17.985342 | 8.685408 | 129.098443 | 62.343696 | 275.700439 | 57.090772 | 100.000000 | 0.0 | 0.0 | 0.0 | -0.321387 |
| 90-95 | 10.354203 | 1.292340 | 12.481311 | 8.241434 | 79.595059 | -298.772462 | -3.093551 | -5.418653 | 0.0 | 0.0 | 0.0 | 0.256932 |
| 95-99 | 8.282277 | 0.139051 | 1.678898 | 7.283324 | 87.938673 | 318.774347 | 2.640177 | 4.624526 | 0.0 | 0.0 | 0.0 | -0.070544 |
| Top 1% | 2.071544 | 0.000000 | 0.000000 | 1.827625 | 88.225265 | 34920.477840 | 72.339316 | 126.709297 | 0.0 | 0.0 | 0.0 | -4.338110 |
TaxBrain comes with two (and counting) built in plots as well
differences_plot(tb_static, 'combined', figsize=(10, 8));
distribution_plot(tb_static, 2021, figsize=(10, 8));
You can run a partial-equlibrium dynamic simulation by initiating the TaxBrian instance exactly as you would for the static reform, but with your behavioral assumptions specified
tb_dynamic = TaxBrain(start_year, end_year, microdata="CPS", reform=reform_url,
behavior={"sub": 0.25})
tb_dynamic.run()
Once that finishes running, we can produce the same weighted total table as we did with the static run.
print("Partial Equilibrium - Combined Tax Liability")
tb_dynamic.weighted_totals("combined")
Partial Equilibrium - Combined Tax Liability
| 2021 | 2022 | 2023 | 2024 | 2025 | 2026 | 2027 | 2028 | 2029 | 2030 | |
|---|---|---|---|---|---|---|---|---|---|---|
| Base | 2.269727e+12 | 3.067767e+12 | 3.246623e+12 | 3.395728e+12 | 3.575938e+12 | 3.979716e+12 | 4.161639e+12 | 4.344151e+12 | 4.543397e+12 | 4.749701e+12 |
| Reform | 2.307804e+12 | 3.119395e+12 | 3.312243e+12 | 3.476774e+12 | 3.673491e+12 | 4.082934e+12 | 4.282357e+12 | 4.483171e+12 | 4.690507e+12 | 4.904587e+12 |
| Difference | 3.807725e+10 | 5.162814e+10 | 6.561941e+10 | 8.104634e+10 | 9.755272e+10 | 1.032176e+11 | 1.207186e+11 | 1.390202e+11 | 1.471100e+11 | 1.548863e+11 |
Or we can produce a distribution table to see details on the effects of the reform.
print("Distribution Table")
tb_dynamic.distribution_table(start_year, "weighted_deciles", "expanded_income", "reform")
Distribution Table
| count | c00100 | count_StandardDed | standard | count_ItemDed | c04470 | c04600 | c04800 | taxbc | c62100 | ... | othertaxes | refund | iitax | payrolltax | combined | ubi | benefit_cost_total | benefit_value_total | expanded_income | aftertax_income | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0-10n | 0.102049 | -7.647952 | 0.018507 | -5.835622 | 0.000000 | 0.000000 | 0.0 | 0.000000 | 0.000000 | -7.764826 | ... | 0.000000 | 0.354646 | -0.354646 | 0.064998 | -0.289648 | 0.0 | 0.799655 | 0.799655 | -6.940854 | -6.651206 |
| 0-10z | 8.391348 | -0.092134 | 8.391348 | 112.514735 | 0.000000 | 0.000000 | 0.0 | 0.000000 | 0.000000 | -0.092134 | ... | 0.000000 | 17.779859 | -17.779859 | 0.000000 | -17.779859 | 0.0 | 0.000000 | 0.000000 | 0.000000 | 17.779859 |
| 0-10p | 12.213289 | 28.286909 | 12.207105 | 167.047672 | 0.006184 | 0.035899 | 0.0 | 0.109175 | 0.003643 | 28.250938 | ... | 0.000000 | 28.404318 | -28.400675 | 3.495389 | -24.905286 | 0.0 | 19.592770 | 19.592770 | 50.389807 | 75.295093 |
| 10-20 | 20.707972 | 207.041657 | 20.230779 | 280.157298 | 0.472337 | 7.619006 | 0.0 | 25.986559 | 2.401345 | 200.049307 | ... | 0.000000 | 68.924569 | -66.523628 | 27.626076 | -38.897552 | 0.0 | 107.290602 | 107.290602 | 323.220845 | 362.118396 |
| 20-30 | 20.707385 | 316.378323 | 19.648591 | 287.017567 | 1.053396 | 18.740597 | 0.0 | 116.225685 | 11.584224 | 299.238036 | ... | 0.000000 | 67.787144 | -56.195886 | 43.577393 | -12.618493 | 0.0 | 258.441980 | 258.441980 | 590.805721 | 603.424214 |
| 30-40 | 20.707088 | 392.642544 | 19.510973 | 297.416383 | 1.187950 | 22.114507 | 0.0 | 184.988094 | 19.315954 | 371.147034 | ... | 0.000000 | 67.101813 | -47.772787 | 53.603619 | 5.830832 | 0.0 | 369.313918 | 369.313918 | 787.711190 | 781.880357 |
| 40-50 | 20.708472 | 550.480439 | 19.029242 | 307.522806 | 1.676905 | 33.455448 | 0.0 | 286.992712 | 30.516102 | 519.344714 | ... | 0.000000 | 78.292375 | -47.767104 | 72.879144 | 25.112041 | 0.0 | 404.535394 | 404.535394 | 994.589197 | 969.477157 |
| 50-60 | 20.707393 | 735.630028 | 18.331876 | 326.411031 | 2.373124 | 49.426182 | 0.0 | 425.521954 | 47.073700 | 691.665427 | ... | 0.000000 | 93.114874 | -46.034925 | 95.833407 | 49.798482 | 0.0 | 466.945896 | 466.945896 | 1258.080439 | 1208.281958 |
| 60-70 | 20.707515 | 959.511949 | 17.811997 | 357.787922 | 2.893239 | 62.644720 | 0.0 | 606.027111 | 72.245168 | 905.890709 | ... | 0.000000 | 112.380798 | -40.126138 | 122.146512 | 82.020374 | 0.0 | 578.376040 | 578.376040 | 1603.049167 | 1521.028793 |
| 70-80 | 20.707197 | 1365.708941 | 16.719155 | 373.852288 | 3.982463 | 98.050853 | 0.0 | 941.858952 | 119.044547 | 1288.034856 | ... | 0.000000 | 121.724695 | -2.658703 | 172.536431 | 169.877728 | 0.0 | 634.000703 | 634.000703 | 2085.339711 | 1915.461983 |
| 80-90 | 20.707626 | 2093.482677 | 14.553142 | 349.813909 | 6.152092 | 179.989249 | 0.0 | 1578.829459 | 218.697073 | 1951.769562 | ... | 0.004426 | 125.344051 | 93.372100 | 257.189635 | 350.561734 | 0.0 | 640.771616 | 640.771616 | 2872.252958 | 2521.691223 |
| 90-100 | 20.708041 | 6246.236344 | 9.453390 | 234.654561 | 11.254651 | 386.830268 | 0.0 | 5608.201771 | 1197.758029 | 5942.795098 | ... | 13.805670 | 47.686797 | 1165.232430 | 553.861208 | 1719.093638 | 0.0 | 512.904614 | 512.904614 | 6987.435086 | 5268.341448 |
| ALL | 207.075375 | 12887.659726 | 175.906106 | 3088.360551 | 31.052341 | 858.906729 | 0.0 | 9774.741472 | 1718.639786 | 12190.328720 | ... | 13.810095 | 828.895941 | 904.990179 | 1402.813813 | 2307.803992 | 0.0 | 3992.973188 | 3992.973188 | 17545.933268 | 15238.129276 |
| 90-95 | 10.353704 | 1667.324398 | 5.676570 | 140.884353 | 4.677134 | 148.216845 | 0.0 | 1374.670598 | 220.454555 | 1553.995778 | ... | 0.030698 | 34.014434 | 186.471142 | 189.172816 | 375.643958 | 0.0 | 284.797859 | 284.797859 | 2039.255590 | 1663.611631 |
| 95-99 | 8.283571 | 2341.021395 | 3.296412 | 82.219564 | 4.987159 | 170.622619 | 0.0 | 2075.284086 | 397.064195 | 2203.591113 | ... | 1.300444 | 13.626819 | 384.737819 | 212.266995 | 597.004815 | 0.0 | 192.019886 | 192.019886 | 2627.393087 | 2030.388273 |
| Top 1% | 2.070766 | 2237.890550 | 0.480409 | 11.550644 | 1.590357 | 67.990803 | 0.0 | 2158.247087 | 580.239279 | 2185.208206 | ... | 12.474527 | 0.045543 | 594.023468 | 152.421397 | 746.444865 | 0.0 | 36.086868 | 36.086868 | 2320.786409 | 1574.341544 |
16 rows × 24 columns
Dynamic Reform with Corporate Income Tax Incidence#
Now we simulate a dynamic revenue estimate while accounting for the incidence of a corporate income tax change.
# Corporate revenue estimate
corp_rev = [5_000_000_000] * (end_year - start_year + 1)
incidence_assumptions = {
"Incidence": { # long-run incidence of corporate tax
"Labor share": 0.5,
"Shareholder share": 0.4,
"All capital share": 0.1,
},
"Long run years": 10, # number of years to reach long-run incidence
}
tb_dynamic = TaxBrain(start_year, end_year, microdata="CPS", reform=reform_url,
behavior={"sub": 0.25},
corp_revenue=corp_rev,
corp_incidence_assumptions=incidence_assumptions)
tb_dynamic.run()
print("Partial Equilibrium - Combined Tax Liability")
tb_dynamic.weighted_totals("combined")
Partial Equilibrium - Combined Tax Liability
| 2021 | 2022 | 2023 | 2024 | 2025 | 2026 | 2027 | 2028 | 2029 | 2030 | |
|---|---|---|---|---|---|---|---|---|---|---|
| Base | 2.269727e+12 | 3.067767e+12 | 3.246623e+12 | 3.395728e+12 | 3.575938e+12 | 3.979716e+12 | 4.161639e+12 | 4.344151e+12 | 4.543397e+12 | 4.749701e+12 |
| Reform | 2.309448e+12 | 3.121009e+12 | 3.313788e+12 | 3.478349e+12 | 3.675071e+12 | 4.084663e+12 | 4.284063e+12 | 4.484873e+12 | 4.692225e+12 | 4.906309e+12 |
| Difference | 3.972134e+10 | 5.324191e+10 | 6.716455e+10 | 8.262134e+10 | 9.913307e+10 | 1.049468e+11 | 1.224239e+11 | 1.407223e+11 | 1.488282e+11 | 1.566081e+11 |