Tax-Calculator has been designed using object-oriented programming (OOP) principles. There are seven classes and a collection of global utility functions, but most Python programming involves using only a few methods in three classes.
Typical Tax-Calculator usage involves creating two Calculator class objects: both containing the same sample of filing units (that is, Records class object), but each containing a different tax policy (that is, Policy class object). The idea is to compare the calculated tax liabilities of the sample units under the two different tax policies, one of which is usually current-law policy and the other is a tax reform of interest.
rec→ Records class object.
Records()when containing IRS-SOI-PUF-derived filing-unit data or created by
Records.cps_constructor()when containing CPS-derived filing-unit data.
Policyclass object containing parameters that characterize current-law policy.
Policyclass object containing parameters that characterize a tax reform.
Created using a Python dictionary
refdictrepresenting the reform by using the
implement_reform(refdict)method on a
Policyobject created by
Policy(). Or created using a JSON file
filenamerepresenting the reform by using the
implement_reform(Policy.read_json_reform(filename))method on a
Policyobject created by
calc_clp→ Calculator class object for current-law policy.
calc_ref→ Calculator class object for reform policy.
calc_all()→ Calculator class method that computes tax liability (and many intermediate variables such as AGI) for each filing-unit.
itax_clp→ Variable containing aggregate income tax liability under current-law policy.
weighted_total('iitax')method called on
diff_table→ Pandas DataFrame object containing reform-minus-current-law difference table for income tax liability by expanded-income deciles.
calc_clp.difference_table(calc_ref, 'weighted_deciles', 'iitax')method called after
calc_all()has been called on both Calculator objects.
For examples of Python scripts that use these classes and methods, see Recipes.
For detailed documentation and source code for these three classes, see:
Tax-Calculator contains eight basic classes, and a collection of global utility functions, that together provide the full range of Tax-Calculator capabilities. Here is a description of their role in Tax-Calculator and a link to each the detailed documentation and source code for each class and all its methods.
Data→ Contains basic logic for manipulating cross-sectional data that need to have growth factors and sample weights to age the data to years after the data start year.
Documentation and source code are in data.py.
Records→ Derived from
Dataand contains attributes of each tax filing unit.
Documentation and source code are in records.py.
GrowFactors→ Contains CBO-derived baseline annual growth factors that are used to specify price inflation and wage growth rates in the
Policyclass object and to specify annual growth factors that are applied to monetary attributes of the filing units in the
Documentation and source code are in growfactors.py.
Parameters→ Contains basic value extrapolation, revision, and validation logic for time-varying parameters that can be any of four types and can be either scalar-valued or vector-valued.
Documentation and source code are in parameters.py.
Policy→ Derived from
Parametersand contains tax policy parameters.
Documentation and source code are in policy.py.
GrowDiff→ Derived from
Parametersand contains differences from CBO-derived baseline growth factors in the
Documentation and source code are in growdiff.py.
Consumption→ Derived from
Parametersand contains parameters related to consumption that are used in the
Documentation and source code are in consumption.py.
Calculator→ Contains a
Policyclass object, a
Recordsclass object, and a
Consumptionclass object, plus functions that contain the logic required to calculate income and payroll tax liability for each filing unit.
Documentation and source code are in calculator.py and in calcfunctions.py.
Documentation and source code for the global utility functions are in utils.py.