`iv_robust()` diagnostics and fixed-effects

[MaelAstruc]

Hello,

Thank you for this package and the helpful functions.

I would like to know why the iv_robust cannot return diagnostics if fixed-effects are specified. I couldn't find any details on it in the documentation, vignettes or issues apart from a short mention in #285 saying that the glance method returns NA.

I imagine that this is due to the demeaning of the outcome, regressors, and instruments which modifies the estimated F statistics, but I wonder if this can be solved with intermediary computations or if the two of them are inherently incompatible.

If a solution exists, I would gladly try to implement it and submit a PR.

Best,
Mael Astruc

[ThijsBusschots]

Related to this issue: I would like to use functions akin to augment() from the broom package, or cooks.distance(). However, both ar not applicable to an object of class iv_robust.

Perhaps anyone is aware of other functions that can be applied to retrieve the cooks distance from an iv_robust object, without having to code the function myself. I would be happy to hear.