HoweWeSolveIndShockConsumerType uses but does not define acronym EGM

[sbenthall]

"EGM" stands for "endogenous grid method"

notebooks/HoweWeSolveIndShockConsumerType.py uses this acronym without defining it.
One has to look around in other notebooks to find a link to the canonical resource.

http://www.econ2.jhu.edu/people/ccarroll/EndogenousGridpoints.pdf

[sbenthall]

EGM is clearly quite important.
If understanding of it is assumed in the audience of all DemARKs, then perhaps that misses an opportunity to use HARK/DemARKS as a way to explicate EGM.

If I understand correctly, EGM is an algorithm.

Latex has a nice package for elucidating algorithms:
https://en.wikibooks.org/wiki/LaTeX/Algorithms

It would be nice if some explanation of EGM was available either in the documentation for HARK or as a DemARK.
The audience ("pathway") for this would be those trained in computer science or programming coming to the libraries from "the internet", not through a classroom.

[sbenthall]

I wonder if EGM exists somewhere in the HARK library as a method.

(As opposed to being reimplemented in many special cases across various model-specific solution code.)

[llorracc]

A quick solution would be to point to the QuantEcon treatment of it:

https://julia.quantecon.org/dynamic_programming/egm_policy_iter.html#Overview

or at

https://llorracc.github.io/SolvingMicroDSOPs/#The-Method-of-Endogenous-Gridpoints

[sbenthall]

Ah, thank you.
The QuantEcon presentation of it is, for me, very clean.
For example, I wouldn't have known EGM was an algorithm for policy iteration without the language in it.