Writing Python Applications that Import the USW-Tax-Analyzer Package
Here we show how to use the uswtaxanalyzer package in a Python
application that computes the change in aggregate wealth tax liability
caused by the reform discussed in the CLI documentation section
entitled Comparing No-Avoidance
Results. The Python
application produces the same results, but only after some time spent
writing the Python code. Python applications that undertake
non-static analysis (such as avoidance responses to the reform) are
considerably more complex than the example presented here.
Here is the Python code, which has been composed in a text editor and
saved in the app.py file:
$ tacat app.py
"""
This USW-Tax-Analyzer application produces the same 2019 aggregate wealth
tax liability estimates as produced by the CLI example in the documentation
under the assumption of no avoidance response to the reform.
"""
import uswtaxanalyzer as ta
cyr = 2019
pol = ta.Policy() # represents current-law policy (that is, no wealth tax)
rec = ta.Records() # uses built-in data, weights, and default growth factors
# create Calculator object for current-law policy and calculate tax liability
calc1 = ta.Calculator(policy=pol, records=rec)
# the calc1 object contains private copies of pol and rec,
# so we are free to use pol and rec however we want below
calc1.advance_to_year(cyr)
calc1.calculate()
dframe1 = calc1.dataframe(['wtax', 'weight'])
del calc1
wtax1 = (dframe1.wtax * dframe1.weight).sum()
# create Calculator object for reform policy and calculate tax liability
# specify Warren reform used in the CLI example as a Python dictionary
# rather than as the JSON text in the ew.json file used by the CLI
ref = {
# 1st is a zero rate bracket for wealth up to $50 million
'brak1': {2019: 50e6}, 'rate1': {2019: 0.00},
# 2nd is a 2% rate bracket from $50 million to $1 billion
'brak2': {2019: 1e9}, 'rate2': {2019: 0.02},
# 3rd is a 3% rate bracket for wealth over $1 billion
'brak3': {2019: 9e99}, 'rate3': {2019: 0.03}
}
pol.implement_reform(ref)
calc2 = ta.Calculator(policy=pol, records=rec)
calc2.advance_to_year(cyr)
calc2.calculate()
dframe2 = calc2.dataframe(['wtax', 'weight'])
del calc2
wtax2 = (dframe2.wtax * dframe2.weight).sum()
ntaxed = dframe2[dframe2.wtax > 0].weight.sum()
print(f'{cyr} CurLaw WTax Liability ($b) = {(wtax1 * 1e-9):.1f}')
print(f'{cyr} Reform WTax Liability ($b) = {(wtax2 * 1e-9):.1f}')
print(f'Number of households taxed (#) = {ntaxed:.0f}')
When this application is executed at the command prompt, it produces these results:
$ python app.py 2019 CurLaw WTax Liability ($b) = 0.0 2019 Reform WTax Liability ($b) = 254.7 Number of households taxed (#) = 93895
These results are the same as generated by this CLI run and dump output tabulation:
$ uswta tax.csv 2019 ew.json && awk -f opa.awk tax-19-ew.csv INPUT data are extrapolated to 2019 using default growth factors under CurLaw INPUT data are extrapolated to 2019 using default growth factors under REFORM Calculations for 2019 use no response elasticities Write minimal REFORM dump output file ./tax-19-ew.csv for 2019 Write table output file ./tax-19-ew.tab for 2019 Write REFORM parameter output file ./tax-19-ew.par for 2019 Run time is 0.1 seconds BITSS-OPA-style results for 2019 total_wealth_tax_($b)= 254.7 total_households_(#m)= 156.436 taxed_households_(#)= 93895 percent_hhs_taxed(%)= 0.06