Report estimates from Heckman AND margins command

[Hanna Lindstrom]

Dear all,
I am using Stata 14 and have some questions regarding the commands -Heckman-, -margins- and -outreg2-.

I am trying to produce tables of my regression results using outreg2.
I am running Stata's Heckman and margins command, using following code:

 heckman y1 `heckvarlist', ///
select(sel=`probvarlist') vce(cluster id) first

margins, dydx(`heckvarlist') predict(ycond) atmean post
outreg2 using "myfile.xls", append ctitle(dy/dx Second stage pooled Heckman) dec(3)

However, I would like my output table to include estimates from the original regression, and not just the marginal effects. I read up on Outreg2 and understand that it's possible to create local macros after running -Heckman- in order to store estimates of e.g. lambda and rho, and include it in the outreg2 output using the addstat command, according to:

heckman y1 `heckvarlist', ///
select(sel=`probvarlist') vce(cluster id) first

local Lambda=e(lambda)
local Selambda=e(selambda)
local rho=e(rho)

outreg2 using "myfile.xls", adds(Lambda, `Lambda', Selambda, `Selambda', rho, `rho') append ctitle(dy/dx Second stage pooled Heckman) dec(3)

Which makes the output look like:

 VARIABLES    dy/dx Second stage pooled Heckman    
x1    0.712***    
      (0.029)    
x2    0.239***    
      (0.067)    
x3   0.201***    
     (0.063)    
x4    -0.619***    
    (0.158)    
x5    0.362***    
    (0.140)    
x6    0.028    
    (0.033)    
        
Observations    3,480      
Lambda         -4.405    
Selambda        0.512    
rho            -0.615        
Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

However, I would also like to display the standard error of rho, which is displayed at the bottom of the output table after Heckman, i.e.:

       _cons |   2.610507   .8621316     3.03   0.002     .9207605    4.300254
-------------+----------------------------------------------------------------
     /athrho |  -.7173276   .0944943    -7.59   0.000     -.902533   -.5321223
    /lnsigma |   1.968422   .0338867    58.09   0.000     1.902005    2.034839
-------------+----------------------------------------------------------------
         rho |  -.6152512    .058725                      -.717529   -.4870017
       sigma |   7.159369   .2426076                      6.699314    7.651017
      lambda |  -4.404811   .5117772                     -5.407875   -3.401746
------------------------------------------------------------------------------
Wald test of indep. eqns. (rho = 0): chi2(1) =    57.63   Prob > chi2 = 0.0000

In the Stata help file on Heckman, I can't find that estimate stored, though. Does anyone know if it is possible to write a command for including it in the outreg2? I know from the helpfile that the standard error of rho is computed using the delta method, does that help me in any way?

Best regards,

Hanna Lindström