Move calculation of kinderzuschlag_max

[Eric-Sommer]

Current situation

In PR #267, we solved the new calculation for the maximum kinderzuschlag by setting the kinderzuschlag_max in policy_enviroment.py (link). While this works fine, this is not the place one would look for it.

Suggestion

by @mjbloemer :

Why not simply add a new year switch (two functions) there where ..bis_2020 is equal to the parameter and ..since_2021 is a function of the ExMin and KG parameters? Just another level of function composition instead of overwriting parameters?

[hmgaudecker]

The problem is that if it was placed in kinderzuschlag.py, it would become part of the DAG. That's why we put it where it is.

But I think this should actually happen! It is also a good example of the value of setting kinderzuschlag_max to null starting in 2021.

The function could just be (untested)

def kinderzuschlag_max(hh_id, kinderzuschlag_params, kindergeld_params):
    if "kinderzuschlag_max" in kinderzuschlag_params:
        out_scalar = kinderzuschlag_params["kinderzuschlag_max"]
    else:
        params["kinderzuschlag"]["kinderzuschlag_max"] = (
            params["kinderzuschlag"]["exmin"]["regelsatz"]["kinder"]
            + params["kinderzuschlag"]["exmin"]["kosten_der_unterkunft"]["kinder"]
            + params["kinderzuschlag"]["exmin"]["heizkosten"]["kinder"]
        ) / 12 - params["kindergeld"]["kindergeld"][1]

    return pd.Series(out_scalar, index=hh_id)

(stuff under else just taken from current function)

[ChristianZimpelmann]

Shouldn't it rather be:

if "kinderzuschlag_max" not in kinderzuschlag_params:

[Eric-Sommer]

Shouldn't it rather be:

if "kinderzuschlag_max" not in kinderzuschlag_params:

I agree. @mjbloemer is however right in pointing out that this leads to errors before 2004 when there is no [exmin] parameter. So maybe if "kinderzuschlag_max" not in kinderzuschlag_params and "exmin" in kinderzuschlag_params:

[mjbloemer]

My point was not about the exmin parameter, but that there is no Kinderzuschlag prior to 2005.

[hmgaudecker]

Shouldn't it rather be:

if "kinderzuschlag_max" not in kinderzuschlag_params:

Thanks! Almost. I fixed it above directly

[hmgaudecker]

My point was not about the exmin parameter, but that there is no Kinderzuschlag prior to 2005.

And you are completely right, of course. You're just too fast for me :wink:

I added that to GEP-04, we'll need to find a good way of handling such cases at the function level. The function kinderzuschlag_max should not be part of the DAG prior to the existence of kinderzuschlag.

See here, docs take a while to build, we probably committed too much today.

[MImmesberger]

Only part missing is #789. Hence, I'll close this one as completed.