How to calculate Elterngeld#

Elterngeld (parental leave benefit) provides financial support to parents during the first months after the birth of a child. It replaces a portion of the parent’s pre-birth income — typically about 67% of net income. This net income is approximated using simplified calculation rules. This document explains how to calculate Elterngeld based on historical income data for a given household.

Overview#

Elterngeld calculation generally involves three main steps:

  1. Calculate the parent’s average monthly gross income in the 12 months before the child’s birth.
    Use this as input to compute
    ("elterngeld", "mean_nettoeinkommen_für_bemessungsgrundlage_nach_geburt_m")
    applying the policy environment from the year before the child’s birth.
  2. Calculate the parent’s gross income in the tax year prior to the child’s birth.
    Use this to compute
    ("einkommensteuer", "zu_versteuerndes_einkommen_y_sn")
    using the policy environment of the birth year.

  3. Use the results from steps 1 and 2 as inputs to calculate the final Elterngeld amount ("elterngeld", "betrag_m") using the policy environment of the birth year.

Example#

Consider a child born in July 2024. The parent had a monthly gross wage of €2,000 in the 12 months before birth, and a yearly gross income of €26,000 in 2023.

An example Elterngeld application form from the state of Bavaria can be found here.

Step 1: Calculate Average Monthly Gross Income Before Birth#

If income from dependent employment is the only source of income, the basis for the Elterngeld calculation is the average monthly gross income in the 12 months before the birth of a child. In our case, this would be July 2023 until June 2024. If other income sources exist, the basis is the average monthly gross income in the calendar year before the birth of the child, 2023 in our example. We will stick to the first case here because it is the only thing that GETTSIM will handle directly right now.

Note: The Elterngeld implementation in GETTSIM does currently not consider income from self-employment or other sources. See this issue.

So let’s assume that income from dependent employment is the only source of income. If you have only annual income data (and thus cannot compute gross income 12 months before birth exactly), you must choose an approximation. In most cases, a good approximation is the income of the year before birth.

[1]:

import numpy as np

from gettsim import InputData, MainTarget, TTTargets, main

The correct point in time for the data, i.e., the relevant state variables for determining the net wage, are an interesting case. Many of those may change over time. The relevant part is:

Die entsprechenden Abzugsmerkmale werden den Lohn- oder Gehaltsabrechnungen entnommen. Grundsätzlich sind die Abzugsmerkmale maßgeblich, die in der überwiegenden Zahl der Monate des Bemessungszeitraums vor Geburt des Kindes gegolten haben.

So it is up to the data preparer to find out which tax class (Steuerklasse), church tax (Kirchensteuer), health insurance, etc. to set. The Steuerklasse, for example, is always set on a yearly basis; if it changes with the start of the calendar year a child is born, the date of birth matters. Up to June, the Steuerklasse from the previous calendar year will be relevant; for July or later, the Steuerklasse from the current calendar year will be relevant.

GETTSIM provides an easy way to compute the relevant net wage ("elterngeld", "mean_nettoeinkommen_in_12_monaten_vor_geburt_m") using the target ("elterngeld", "mean_nettoeinkommen_für_bemessungsgrundlage_nach_geburt_m").

We use the policy environment of January 1st of the year before the child was born (§2e Abs. 1 S. 2 BEEG). Note that this is correct regardless of the point in time when the child is born

In our case, this is January 1st, 2023.

Here, we use some example data with ("einnahmen", "bruttolohn_m") being 2,000€ per month.

[2]:

input_data_12_months_before_birth = {
    "alter": np.array([30]),
    "arbeitsstunden_w": np.array([39]),
    "familie": {
        "p_id_elternteil_1": np.array([-1]),
        "p_id_elternteil_2": np.array([-1]),
    },
    "kindergeld": {"in_ausbildung": np.array([False])},
    "lohnsteuer": {"steuerklasse": np.array([1])},
    "p_id": np.array([0]),
    "hh_id": np.array([0]),
    "sozialversicherung": {"pflege": {"beitrag": {"hat_kinder": np.array([False])}}},
    "wohnort_ost_hh": np.array([False]),
    "einnahmen": {
        "bruttolohn_m": np.array([2000]),
    },
}

Now, we calculate the approximated net wage in the 12 months before birth.

[3]:

net_wage_12_months_before_birth = main(
    main_target=MainTarget.results.tree,
    policy_date_str="2023-01-01",
    tt_targets=TTTargets.tree(
        {
            "elterngeld": {
                "mean_nettoeinkommen_für_bemessungsgrundlage_nach_geburt_m": None,
            }
        }
    ),
    input_data=InputData.tree(input_data_12_months_before_birth),
)["elterngeld"]["mean_nettoeinkommen_für_bemessungsgrundlage_nach_geburt_m"]
net_wage_12_months_before_birth

[3]:

array([1455.12])

Step 2: Taxable income in the tax year before birth#

Parental leave benefits are paid out only if the parent’s taxable income in the tax year before the child was born is below a certain threshold. We calculate this using another GETTSIM call.

First, you have to compile all the necessary data to compute the parent’s taxable income in the year before the child was born (here: 2023). Here, we just use some made-up data.

[4]:

input_data_2023 = {
    "alter": np.array([30]),
    "arbeitsstunden_w": np.array([39]),
    "behinderungsgrad": np.array([0]),
    "einkommensteuer": {
        "abzüge": {
            "beitrag_private_rentenversicherung_m": np.array([0]),
            "kinderbetreuungskosten_m": np.array([0]),
            "p_id_kinderbetreuungskostenträger": np.array([0]),
        },
        "einkünfte": {
            "aus_forst_und_landwirtschaft": {"betrag_y": np.array([0])},
            "aus_gewerbebetrieb": {"betrag_y": np.array([0])},
            "aus_selbstständiger_arbeit": {"betrag_y": np.array([0])},
            "aus_vermietung_und_verpachtung": {"betrag_y": np.array([0])},
            "ist_hauptberuflich_selbstständig": np.array([False]),
            "sonstige": {
                "alle_weiteren_y": np.array([0]),
                "rente": {"betrag_m": np.array([0])},
            },
        },
        "gemeinsam_veranlagt": np.array([False]),
    },
    "einnahmen": {
        "bruttolohn_y": np.array([26000]),
        "kapitalerträge_y": np.array([0]),
    },
    "familie": {
        "alleinerziehend": np.array([False]),
        "p_id_ehepartner": np.array([-1]),
        "p_id_elternteil_1": np.array([-1]),
        "p_id_elternteil_2": np.array([-1]),
    },
    "geburtsjahr": np.array([1993]),
    "grundsicherung": {"im_alter": {"betrag_m_eg": np.array([0])}},
    "kindergeld": {
        "in_ausbildung": np.array([False]),
        "p_id_empfänger": np.array([-1]),
    },
    "p_id": np.array([1]),
    "sozialversicherung": {
        "kranken": {
            "beitrag": {
                "privat_versichert": np.array([False]),
                "bemessungsgrundlage_rente_m": np.array([0]),
            }
        },
        "pflege": {"beitrag": {"hat_kinder": np.array([False])}},
    },
    "wohnort_ost_hh": np.array([False]),
}

Now, we calculate the taxable income in 2023.

[5]:

taxable_income_2023 = main(
    main_target=MainTarget.results.tree,
    policy_date_str="2023-01-01",
    tt_targets=TTTargets.tree(
        {
            "einkommensteuer": {
                "zu_versteuerndes_einkommen_y_sn": None,
            }
        }
    ),
    input_data=InputData.tree(input_data_2023),
    include_warn_nodes=False,
)["einkommensteuer"]["zu_versteuerndes_einkommen_y_sn"]
taxable_income_2023

[5]:

array([19806.])

Step 3: Compute Elterngeld#

Finally, we can compute Elterngeld using the target ("elterngeld", "betrag_m") and the input data generated in step 1 and 2.

The reason for not doing this in one GETTSIM call (together with step 1 and 2) is that

  1. we have to use different policy environment dates for the two steps, and

  2. other state variables of the household (e.g. place of residence) might change after birth of the child.

First, we create the new data set after birth.

[6]:

input_data_after_birth = {
    "alter": np.array([31, 0]),
    "alter_monate": np.array([372, 0]),
    "arbeitsstunden_w": np.array([0, 0]),
    "bürgergeld": {"p_id_einstandspartner": np.array([-1, -1])},
    "einnahmen": {"bruttolohn_m": np.array([0, 0])},
    "elterngeld": {
        "bisherige_bezugsmonate": np.array([0, 0]),
        "claimed": np.array([True, False]),
        "mean_nettoeinkommen_in_12_monaten_vor_geburt_m": np.append(
            net_wage_12_months_before_birth, np.array([0])
        ),
        "zu_versteuerndes_einkommen_vorjahr_y_sn": np.append(
            taxable_income_2023, np.array([0])
        ),
    },
    "familie": {
        "alleinerziehend": np.array([True, False]),
        "p_id_elternteil_1": np.array([-1, 0]),
        "p_id_elternteil_2": np.array([-1, -1]),
    },
    "hh_id": np.array([0, 0]),
    "kindergeld": {"in_ausbildung": np.array([False, False])},
    "lohnsteuer": {"steuerklasse": np.array([1, 1])},
    "p_id": np.array([0, 1]),
    "sozialversicherung": {
        "pflege": {"beitrag": {"hat_kinder": np.array([True, False])}}
    },
    "wohnort_ost_hh": np.array([False, False]),
}

Then, we compute Elterngeld using the policy environment of the child’s birthday.

[7]:

main(
    main_target=MainTarget.results.tree,
    policy_date_str="2024-07-01",
    tt_targets=TTTargets.tree(
        {
            "elterngeld": {
                "betrag_m": None,
            }
        }
    ),
    input_data=InputData.tree(input_data_after_birth),
)["elterngeld"]["betrag_m"]

[7]:

array([945.82,   0.  ])