Support cluster argument to feols()

[SebKrantz]

options(fastverse.styling = FALSE)
library(fastverse)
#> -- Attaching packages --------------------------------------- fastverse 0.3.2 --
#> v data.table 1.14.8     v kit        0.0.13
#> v magrittr   2.0.3      v collapse   2.0.7
fastverse_extend(fixest, wildrwolf)
#> -- Attaching extension packages ----------------------------- fastverse 0.3.2 --
#> Warning: package 'fixest' was built under R version 4.3.1
#> v fixest    0.11.2     v wildrwolf 0.6.1
#> -- Conflicts ------------------------------------------ fastverse_conflicts() --
#> x fixest::fdim() masks collapse::fdim()
models = feols(c(vs, am) ~ mpg | cyl, mtcars, cluster = "carb")
rwolf(models = models, param = "mpg", B = 9999)
#>   |                                                                              |                                                                      |   0%
#> Error in formula.character(clustid): invalid formula "carb": not a call

models = feols(c(vs, am) ~ mpg | cyl, mtcars, cluster = ~carb)
rwolf(models = models, param = "mpg", B = 9999)
#>   |                                                                              |                                                                      |   0%
#> Warning: Please note that the seeding behavior for random number generation for
#> `boottest()` has changed with `fwildclusterboot` version 0.13.
#> 
#> It will no longer be possible to exactly reproduce results produced by versions
#> lower than 0.13.
#> 
#> If your prior results were produced under sufficiently many bootstrap
#> iterations, none of your conclusions will change.  For more details about this
#> change, please read the notes in
#> [news.md](https://cran.r-project.org/web/packages/fwildclusterboot/news/news.html).
#> This warning is displayed once per session.
#> Warning: There are only 64 unique draws from the rademacher distribution for 6 bootstrap
#> clusters. Therefore, B = 64 with full enumeration. Consider using webb weights
#> instead. Further, note that under full enumeration and with B = 64 bootstrap
#> draws, only 2^(#clusters - 1) = 32 distinct t-statistics and p-values can be
#> computed. For a more thorough discussion, see Webb `Reworking wild bootstrap
#> based inference for clustered errors` (2013).
#>   |                                                                              |===================================                                   |  50%
#> Warning: There are only 64 unique draws from the rademacher distribution for 6 bootstrap
#> clusters. Therefore, B = 64 with full enumeration. Consider using webb weights
#> instead. Further, note that under full enumeration and with B = 64 bootstrap
#> draws, only 2^(#clusters - 1) = 32 distinct t-statistics and p-values can be
#> computed. For a more thorough discussion, see Webb `Reworking wild bootstrap
#> based inference for clustered errors` (2013).
#>   |                                                                              |======================================================================| 100%
#>   model     Estimate Std. Error    t value    Pr(>|t|) RW Pr(>|t|)
#> 1     1 -0.006000719 0.01009511 -0.5944186   0.5781065   0.9692308
#> 2     2   0.04767382 0.01039968   4.584163 0.005924984   0.2000000

models = feols(c(vs, am) ~ mpg | cyl, mtcars, cluster = as.character(mtcars$carb))
rwolf(models = models, param = "mpg", B = 9999)
#>   |                                                                              |                                                                      |   0%
#> Error in terms.formula(formula, data = data): invalid model formula in ExtractVars

^{Created on 2023-12-12 with reprex v2.0.2}

[s3alfisc]

Thanks for reporting. I have plenty of time off between now and the new year and will make sure to take a look!

[s3alfisc]

Some of this should be fixed now and available on r-universe in the next hour. CRAN might take me a little longer 😅

models = feols(c(vs, am) ~ mpg | cyl, mtcars, cluster = as.character(mtcars$carb))
rwolf(models = models, param = "mpg", B = 9999)

This one I have not fixed, as boottest() does not allow for vector-valued input. I'll work towards a nicer error message in
this case.

[SebKrantz]

Thanks! I see you want to keep it open, otherwise I would have closed it now.