# st: Question on Simulating Standard Errors(S.E)

 From Asgar Khademvatani To statalist@hsphsun2.harvard.edu Subject st: Question on Simulating Standard Errors(S.E) Date Sat, 08 Dec 2007 15:04:32 -0700

Dear All,

Basically, I have estimated a system of the cost and labor-share equation using ISURE. I am using time-series of a particular U.S. industry stretched over 1958-2000 for 43 years. Then I have generated a variable called "sve_sg" as this new variable is a function of all RHS variables of the cost function. Then, I have generated the elasticities of this new defined variable "sve_sg" with respect to each factor, as it is function of that factor. For example, If I would like to calculate its elasticity with respect to the factor"l" , in doing so, the first, I generate this elasticity, named, "Isvl_sr_sg" as you see in the following. so that this generated elasticity is a function of some variables that each are as function of the estimated coefficients and RHS variables of the cost function as follows;

Isvl_sr_sg = (d2cel_sg)*((pl_sg)/(sve_sg))

My target is the simulating the standard errors(S.E) of such elasticity over 43 years. To do so, my procedure is that I re-run the ISURE model and then I am using the following "testnl" command as follows:

est replay "ISURE"
testnl [=lvc_sg]: Isvl_sr_sg // lvc_sg is the cost function.

But, after runnign my .do file, I am getting the following error. Can anybody help me out that what is wrong with using "testnl" command?

testnl [=lvc_sg]: Isvl_sr_sg

(1) [ = lvc_sg]: Isvl_sr_sg
r(303);

end of do-file
r(303);

I am not sure whether this procedure is correct to simulate S.E of the Isvl_sr_sg or not. If not so, anybody please can help me on this matter, how to calculate the S.E though any other way?

Thank you

Asgar Kh.