3

Does anyone know about a R package that supports fixed effect, instrumental variable regression like xtivreg in stata (FE IV regression). Yes, I can just include dummy variables but that just gets impossible when the number of groups increases.

Thanks!

landroni
  • 2,902
  • 1
  • 32
  • 39
user2503795
  • 4,035
  • 2
  • 34
  • 49

3 Answers3

4

I can just include dummy variables but that just gets impossible when the number of groups increases

By "impossible," do you mean "computationally impossible"? If so, check out the plm package, which was designed to handle cases that would otherwise be computationally infeasible, and which permits fixed-effects IV.

Start with the plm vignette. It will quickly make clear whether plm is what you're looking for.

Update 2018 December 03: the estimatr package will also do what you want. It's faster and easier to use than the plm package.

user697473
  • 2,165
  • 1
  • 20
  • 47
  • Yes, "computationally impossible" because of the 1000+ coefficients that are estimated in the dummy-variable implementation. I think `plm ` does what I want but it's just soooo slow and I am not sure why. I am now using a test dataset with about 3000 students in 80 school and a simple fixed effect model. Below is some example code. Stata estimates this model nearly instantaneously and `plm` takes forever... `df=plm.data(df, index = "id_school")` `m1=plm(y~x1+x2+x3+x4+x5,data=df, model="within")` – user2503795 Jul 05 '12 at 08:02
  • I don't know why `plm` is so slow in that case. You can skip `plm.data` and just run `plm(y ~ x1 + x2 +x3 + x4 +x5, data = df, model = "within", index = c('id_school')`, but I am not sure that will help much. If it doesn't, all I can think is that you could profile the code. Or maybe swap in a better BLAS library. Some people say that swapping in a different BLAS [doesn't help](http://www.r-bloggers.com/faster-r-through-better-blas/), but it clearly helped me when I was running Windows. See http://prs.ism.ac.jp/~nakama/SurviveGotoBLAS2/ for more detail. – user697473 Jul 05 '12 at 12:16
1

As you may know, for many fixed effects and random effects models {I should mention FE and RE from econometrics and education standpoint since the definitions in statistics are different}, you can create an equivalent SEM (Structural Equation Modeling) model. There are two packages in R that can be used for that purpose: 1)SEM 2) LAVAAN

Another solution is to use SAS. In SAS, you can use Proc GLM which enables you to use "absorb" statement which automatically takes care of the dummies as well as finding (x - xbar) per each observation.

Hope it helps.

Sam
  • 4,357
  • 6
  • 36
  • 60
1

Try the ivreg command from the AER package.

David
  • 265
  • 4
  • 13