Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Error did_imputation Borusyak (2021)

    Hello friends,

    I'm applying the did_imputation method proposed by Borusyak (2021). The command without using covariates is working correctly ( did_imputation y id year D, fe(year) ), however when adding covariates it shows the error: Standard error convergence is not achieved for coefs: __w_tau.
    Try increasing the tolerance, the number of iterations or use the nose option for point estimates without SE.
    convergence not reached

    But I'm not solving it, does anyone have a suggestion? Thankss

  • #2
    Reach out to Borusyak. I don't "know", but it's unlikely that very many people here will know the answer to this since this is a very niche question such that you'd need to know the ins and outs of the command/econometrics in order to comment on it. The only men I know of who may know, is FernandoRios or maybe Daniel PV.


    Either way, have you tried the options Borusyak tells you, changing tolerance or iterations?

    Comment


    • #3
      Hi both of you, I'm not really familiar with this bug, so best option is to mail Borusyak or open an issue in did_imputation repository http://github.com/borusyak/did_imputation

      Comment


      • #4
        Originally posted by Jared Greathouse View Post
        Reach out to Borusyak. I don't "know", but it's unlikely that very many people here will know the answer to this since this is a very niche question such that you'd need to know the ins and outs of the command/econometrics in order to comment on it. The only men I know of who may know, is FernandoRios or maybe Daniel PV.


        Either way, have you tried the options Borusyak tells you, changing tolerance or iterations?
        Hello friend,

        Yes, from what I understand, you increase the iterations with tol(), I did it and it worked, as much as the covariates gave negative signs, for my analysis, it makes little sense.

        Comment


        • #5
          Originally posted by Daniel PV View Post
          Hi both of you, I'm not really familiar with this bug, so best option is to mail Borusyak or open an issue in did_imputation repository http://github.com/borusyak/did_imputation
          Thanks my friend!

          Comment


          • #6
            Hi, I‘m now having the same problem. Did you know how to deal with it? Many thanks!

            Comment

            Working...
            X