HARK
HARK copied to clipboard
econ-ark
Heterogenous Agents Resources & toolKit
The `utilities.CRRAutility` method takes a consumption parameter `c` and a risk aversion factor `gam`. When used in practice, the `gam` function needs to be filled in... this is currently done...
HARK does not work unless it can replicate certain substantive economic results. Currently, some of these tests (in deterministic cases) are intermingled with unit tests and run automatically. For tests...
In a downstream application (SHARKFin), we are running into [a situation](https://github.com/sbenthall/SHARKFin/issues/186) where we need to export a HARK solution to disk memory and import it again later. (This is so...
What should the default values be for these parameters? tail_N=0, tail_bound=None, tail_order=np.e
Given the new attribute `DiscreteDistribution.limit` we can find the minimum and maximum values of support for the distributions. These should be used to find minimum and maximum income realizations in...
Found a tool that seems like a game changer for testing the codebase. https://pytest-xdist.readthedocs.io/en/latest/ This option takes advantage of parallel processing of tests. I'm not sure if it can be...
It would be good to have a class that represents a function with its derivatives. This may exist already outside of HARK and can be brought it.
@gms158 does this make the CRRA utility bottle necks that you showed us faster? I know this breaks all the `Fast` code and I have a good idea as to...
Considering #1121 .... Set $k _t= a_{t-1}$ across all the models. The broader issue is reference to variables outside of a time period. It's awkward that newborns are born before...
Subproblem of #914 see also #1171 Currently grids over state spaces are computed when a model is initialized based parameters handed to the model during construction. We would like this...