Singular matrix and coefficient

[liqizheng0328]

Dear Sir or Madam,

There are two questions I could not find the way to solve it. I would appreciate some feedback on these problems. (1) As is shown in the code and error below, how could I check this problem and solve it. (I have dropped the NA and repeated observation, and it works in stata by reghdfe) `> out1 <- att_gt(yname="Y",

• tname="year",
• idname="idnumber1",
• gname="enteryear",
• xformla=~labor_productivity+lnassets+lnage,
• data=did,
• panel = FALSE) Error in solve.default(crossprod(wols.x.pre.treat, int.cov)/n) : Lapack routine dgesv: system is exactly singular: U[4,4] = 0 In addition: There were 11 warnings (use warnings() to see them)

  out1 <- att_gt(yname="Y",

• tname="year",
• idname="idnumber1",
• gname="enteryear",
• xformla=~labor_productivity+lnassets+operating,
• data=did,
• panel = FALSE) Error in solve.default(crossprod(wols.x.pre.treat, int.cov)/n) : system is computationally singular: reciprocal condition number = 3.27372e-19 In addition: There were 11 warnings (use warnings() to see them)`

(2) How can I output the coefficient of the control variable in this package, as it seems not to provide the argument.

Thank you very much!

[bcallaway11]

The error message is saying that the design matrix isn't invertible at least for some combinations of groups and time periods. Will the code run if you use fewer covariates?

[liqizheng0328]

Thanks for your reply Yes, it works when using fewer covariates.

The error message is saying that the design matrix isn't invertible at least for some combinations of groups and time periods. Will the code run if you use fewer covariates?