save lincom results

[francy pampu]

Hi everyone,
I'd like to if there is the possobility to store somewhere or to send to excel (for example with putexcel) the results of the command lincom.
I used lincom within a loop which produces many and many results so I need to put them toghether.
Is it possible?
Last and least: how can I send to excel the scalar variables?
many tks

[Elan Cohen]

Francy,

This should be enough to get you started.

sysuse auto
reg price mpg
lincom mpg
putexcel A1=(r(df)) using junk.xlsx, replace

[francy pampu]

ok. I understan.
But, what if I want to store confidence intervals and p-values?

[Rich Goldstein]

all of that can be found in r(table):

sysuse auto
regress price mpg foreign
mat li r(table)

and use standard Stata functions to get what you want (e.g., help matrix_functions)

[francy pampu]

OK but lincom doesn't produce r(table). Maybe I have to use lincomest ???

[Rich Goldstein]

sorry for misreading - however, lincom does give you the df (r(df)), the standard error (r(se)) and the estimate (r(estimate)) so you can use those along with the correct distributional function (see help density_functions)

[francy pampu]

Yes... maybe I hve to give up... this is the only solution.
I also tried this code that worked well:

forvalues x=2005/2012 {
quietly xtfrontier TC Q1 Q2 Q3 P1 P2 E Q1Q1 Q1Q2 Q1Q3 Q2Q2 Q2Q3 Q3Q3 P1P1 P1P2 P2P2 EE P1Q1 P1Q2 P1Q3 P2Q1 P2Q2 P2Q3 P1E P2E Q1E Q2E Q3E LNGDPCAP LNMKCAP LNPUBDEBT, tvd cost nonrtolerance
lincomest _b[Q1]+_b[Q2]+_b[Q3]+_b[Q1Q1]*Q1m`x'+_b[Q1Q2]*Q1m`x'+_b[Q1Q3]*Q1m`x'+_b[Q2Q2]*Q2m`x'+_b[Q2Q3]*Q2m`x'+_b[Q3Q3]*Q3m`x'+_b[Q1Q2]*Q2m`x'+_b[Q1Q3]*Q3m`x'+_b[Q2Q3]*Q3m`x'+_b[P1Q1]*P1m`x'+_b[P1Q2]*P1m`x'+_b[P1Q3]*P1m`x'+_b[P2Q1]*P2m`x'+_b[P2Q2]*P2m`x'+_b[P2Q3]*P2m`x'+_b[Q1E]*Em`x'+_b[Q2E]*Em`x'+_b[Q3E]*Em`x'
matrix sy`x' = r(table)
matrix colnames sy`x' = `x'
}
matrix scaleyear = sy2005,sy2006,sy2007,sy2008,sy2009,sy2010,sy2011,s y2012

...but the estimation procedure takes a lot of time every loop...