Child and Dependent Care Credit#

The Child and Dependent Care Tax Credit (CDCC) provides a share of care expenses depending on income.

Examples#

Consider a single parent with $8,000 child care expenses per child (the CDCC’s maximum in 2021, up to two children). Since the CDCC is capped at the lower of head and spouse’s earnings, this example is limited to single parents. In 2020, the CDCC phased out only at one level; in 2021 only, it phased out twice, and in 2022 and beyond, it returns to 2020 law.

from policyengine_us import IndividualSim
import pandas as pd
import plotly.express as px


def make_cdcc(children, year):
    sim = IndividualSim(year=year)
    sim.add_person(name="head", is_tax_unit_head=True)
    members = ["head"]
    for i in range(children):
        child = "child{}".format(i)
        sim.add_person(name=child, age=5)
        members += [child]
    sim.add_tax_unit(
        name="tax_unit",
        members=members,
        tax_unit_childcare_expenses=8_000 * children,
    )
    sim.add_spm_unit(name="spm_unit", members=members)
    sim.add_household(name="household", members=members, state_code="MD")
    sim.vary("employment_income", max=500_000)
    return pd.DataFrame(
        dict(
            employment_income=sim.calc("employment_income")[0],
            cdcc=sim.calc("cdcc")[0].round(),
            cdcc_mtr=-sim.deriv(
                "cdcc",
                "employment_income",
                wrt_target="head",
            ),
            children=children,
            year=year,
        )
    )


# Make a table of CDCC amounts for different numbers of adults and children.
l = []
for year in [2020, 2021, 2022]:
    for children in range(1, 3):
        l.append(make_cdcc(children, year))

df = pd.concat(l)

LABELS = dict(
    employment_income="Employment income",
    cdcc="CDCC",
    cdcc_mtr="CDCC Marginal Tax Rate",
    children="Children",
    year="Year",
)

fig = px.line(
    df,
    "employment_income",
    "cdcc",
    color="children",
    animation_frame="year",
    labels=LABELS,
    title="Child and Dependent Care Credit",
)
fig.update_layout(
    xaxis_tickformat="$,",
    yaxis_tickformat="$,",
    yaxis_range=[0, df.cdcc.max() * 1.05],
)
fig.show()

The credit phases in at 50% of income and then phases out in a stepwise fashion that produces many small cliffs (infinite marginal tax rates).

fig = px.line(
    df,
    "employment_income",
    "cdcc_mtr",
    color="children",
    animation_frame="year",
    labels=LABELS,
    title="CDCC marginal tax rate",
)
fig.update_layout(
    xaxis_tickformat="$,", yaxis_tickformat=".1%", yaxis_range=[-1, 1]
)
fig.show()

Budgetary impact#

Applying 2022 rules to the 2020 Current Population Survey, PolicyEngine US estimates that the CDCC pays out $4.0 billion.

from policyengine_us import Microsimulation

sim = Microsimulation(dataset_year=2020)

sim.calc("cdcc", period=2022).sum() / 1e6
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
Cell In[3], line 3
      1 from policyengine_us import Microsimulation
----> 3 sim = Microsimulation(dataset_year=2020)
      5 sim.calc("cdcc", period=2022).sum() / 1e6

File ~/work/policyengine-us/policyengine-us/policyengine_us/system.py:115, in Microsimulation.__init__(self, *args, **kwargs)
    114 def __init__(self, *args, **kwargs):
--> 115     super().__init__(*args, **kwargs)
    117     reform = create_structural_reforms_from_parameters(
    118         self.tax_benefit_system.parameters, year_start
    119     )
    120     if reform is not None:

TypeError: __init__() got an unexpected keyword argument 'dataset_year'

However, since it interacts with other programs, we need to compare total income after repealing the program to estimate the true budgetary impact. Applying this method yields a lower $3.8 billion.

from policyengine_us.model_api import *


class ignore_reported(Reform):
    def apply(self):
        self.neutralize_variable("spm_unit_net_income_reported")


class neutralize_cdcc(Reform):
    def apply(self):
        self.neutralize_variable("cdcc")


sim = Microsimulation(ignore_reported, year=2020)
sim_no_cdcc = Microsimulation((ignore_reported, neutralize_cdcc), year=2020)

(
    sim.calc("spm_unit_net_income", period=2022).sum()
    - sim_no_cdcc.calc("spm_unit_net_income", period=2022).sum()
) / 1e6
3769.459399326172