Luca Bittarello

@MartinStancsicsQC is looking into it in the context of [this PR](https://github.com/Quantco/tabmat/pull/267) in `tabmat`. :)

Since we're doing a breaking release anyway, would it be possible to rearrange the arguments of [`coef_table`](https://github.com/Quantco/glum/blob/e342311d0d497159527bf6572f9cb1d4a6119c75/src/glum/_glm.py#L1343)? In particular, to move `confidence_level` down, closer to `robust` and friends. In most...

> The first UserWarning seems like a false alarm: the GLM is unpenalized! Note that `glum` defaults to a regularised GLM ([ref](https://github.com/Quantco/glum/blob/d4b7b5d37f0bb6571fa1293740aabfade1f39e28/src/glum/_glm.py#L3040)). As @jtilly mentioned in #730, it isn't a...

> The NotImplementedError comes unexpected. That's a shortcoming in `tabmat`: it gags when all types are integers. If the data frame contains any floating columns, the conversion succeeds. It should...

Fix coming as part of Glum 3.

The daily tests are still failing, but for different reasons: https://github.com/Quantco/glum/actions/runs/9153751283/job/25163209834

I define the main types [here](https://github.com/lbittarello/Microeconometrics.jl/blob/master/src/inference/corr.jl). [This file](https://github.com/lbittarello/Microeconometrics.jl/blob/master/src/inference/vcov.jl) contains the programs to compute the variance matrix of one– and two-stage parametric estimators.