option vce() not allowed

[Maria Iocco]

Hello everyone,


I having some trouble with some regressions lately. I have a dataset with 1400 observations. Cross-sectional. Stata version 16, updated.

* Example generated by -dataex-. For more info, type help dataex
clear
input float(var_1 var_2) int iv float weight
1.0869565  -.6574075 166  .02777778
 1.195652  .11661406 102 .015384615
 1.246377 -1.9730076 175 .015151516
1.2608695  .53105706 184 .037037037
1.2826087  -.6574075 166  .02777778
1.2971015  -.1605494 204 .012658228
 1.326087  .11661406 102 .015384615
 1.347826  .11661406 102 .015384615
 1.362319  -.6574075 166  .02777778
1.3768116  .52166766 121  .01234568
end

I run this command from the do-file:

use "project_data.dta", clear

reg var_1 var_2  [aweight=weight], vce(cluster iv)

and I get the error
"option vce() not allowed
r(198);

"


I went and copied the same command into the command window, and I got the same error. The weird thing is that if I type there

reg var_1 var_2 [aweight=weight]

then

reg var_1 var_2 [aweight=weight], robust

then reg var_1 var_2 [aweight=weight], vce(robust)

and finally

reg var_1 var_2 [aweight=weight], vce(cluster iv)

it does work. This is the output

. reg var_1 var_2 [aweight=weight], vce(cluster iv)
(sum of wgt is 23.80889513716102)

Linear regression Number of obs = 1,261
F(1, 26) = 5.29
Prob > F = 0.0297
R-squared = 0.0394
Root MSE = .47606

(Std. err. adjusted for 27 clusters in iv)
------------------------------------------------------------------------------
| Robust
var_1 | Coefficient std. err. t P>|t| [95% conf. interval]
-------------+----------------------------------------------------------------
var_2 | .0969326 .0421414 2.30 0.030 .0103098 .1835554
_cons | 2.366665 .0478555 49.45 0.000 2.268297 2.465034
------------------------------------------------------------------------------

.


Do you have any idea why?

[Carlo Lazzaro]

Maria:
it works at my end:

. input float(var_1 var_2) int iv float weight

         var_1      var_2        iv     weight
  1. 
. 1.0869565  -.6574075 166  .02777778
  2. 
.  1.195652  .11661406 102 .015384615
  3. 
.  1.246377 -1.9730076 175 .015151516
  4. 
. 1.2608695  .53105706 184 .037037037
  5. 
. 1.2826087  -.6574075 166  .02777778
  6. 
. 1.2971015  -.1605494 204 .012658228
  7. 
.  1.326087  .11661406 102 .015384615
  8. 
.  1.347826  .11661406 102 .015384615
  9. 
.  1.362319  -.6574075 166  .02777778
 10. 
. 1.3768116  .52166766 121  .01234568
 11. 
. end

. reg var_1 var_2 [aweight=weight], vce(cluster iv)
(sum of wgt is .2066796468570828)

Linear regression                               Number of obs     =         10
                                                F(1, 5)           =       2.38
                                                Prob > F          =     0.1837
                                                R-squared         =     0.0462
                                                Root MSE          =     .09486

                                     (Std. err. adjusted for 6 clusters in iv)
------------------------------------------------------------------------------
             |               Robust
       var_1 | Coefficient  std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
       var_2 |   .0275348   .0178552     1.54   0.184    -.0183635    .0734331
       _cons |   1.275964   .0141159    90.39   0.000     1.239678     1.31225
------------------------------------------------------------------------------