euromod.eu_euromod.calibrate

Calibrating a parametric tax-benefit function on the EU model.

A tax-benefit system has many instruments. Macroeconomic models need it in a compact form: a function from a household’s gross income to the taxes it pays net of the benefits it receives. This package estimates such a function from a EUROMOD simulation, for a baseline or for a reform.

It works in two steps, one module each:

data_preparation

Runs a system of the model and builds the household sample: one row per household, with gross and disposable income in multiples of mean income.

estimation

Fits the function of Heathcote, Storesletten and Violante, net tax rate = 1 - (1 - lambda) * income**(-tau), and measures how well it fits.

descriptives describes the sample, weighted holds the weighted statistics, and estimates saves and loads estimates. The estimates for the 27 EU member states and the euro area, policy years 2022 to 2025, ship with the package:

from euromod.eu_euromod.calibrate import load_estimates, schedule

portugal = schedule(load_estimates(), "PT_2022")
portugal.net_tax_rate(1.0)            # 0.1925, the rate at mean income
portugal.disposable_income(2.0)       # 1.4424, in multiples of mean income

The functions need a path to the model and a path to the folder with the input data, and the name of a system:

from euromod.eu_euromod.calibrate import prepare_sample, fit_hsv

baseline = prepare_sample(MODEL, DATA, "PT_2022")
fit_hsv(baseline)                     # {'lambda': 0.1925, 'tau': 0.1631, ...}

A reform is a counterfactual system of the model, or a baseline system with some constants overridden:

reform = prepare_sample(MODEL, DATA, "PT_2022_ref")
reform = prepare_sample(MODEL, DATA, "PT_2022", constants={"$tin00_rate9": 0.58})

This package is independent of euromod.calibrate, the connector’s general engine for fitting arbitrary functional forms. to_calibration() turns a result of this package into an object of that engine.

Below are listed the main public classes of the euromod.eu_euromod.calibrate module.

Function

calibrate_hsv(model_path, data_path, systems, *, household_types, pooled, lambda_fixed, constants, cache_dir, **sample_options)

Estimate the HSV schedule for several systems and household types in one call.

Functions

euromod.eu_euromod.calibrate.calibrate_hsv(model_path, data_path, systems, *, household_types=HOUSEHOLD_TYPES, pooled=None, lambda_fixed=None, constants=None, cache_dir=None, **sample_options)

Estimate the HSV schedule for several systems and household types in one call.

Parameters:

  • model_path (str) – Folder of the EUROMOD model.

  • data_path (str) – Folder with the input datasets.

  • systems (list of str) – System names, e.g. ["PT_2022"] or euromod.eu_euromod.eu27_systems() (2022).

  • household_types (dict) – Conditions by name; default HOUSEHOLD_TYPES.

  • pooled (dict, optional) – Groups of countries to estimate on a pooled sample as well, e.g. {"EA_2022": euro_area(2022)}. Every country of a group must be among systems.

  • lambda_fixed (float, optional) – Hold the level at this value and estimate only tau.

  • constants – As in simulate(). constants apply to every system.

  • cache_dir – As in simulate(). constants apply to every system.

  • **sample_options – As in household_sample().

Returns:

One row per system (or pooled group) and household type, with the columns of fit_hsv() and fit_diagnostics(). Keep it with save_estimates().

Return type:

pandas.DataFrame

Example

>>> from euromod.eu_euromod import euro_area, eu27_systems
>>> table = calibrate_hsv(MODEL, DATA, eu27_systems(2022),
...                       pooled={"EA_2022": euro_area(2022)})
>>> table.loc[("PT_2022", "all"), ["lambda", "tau"]]