Equation has been stored but it is not found to perform test

[Eduardo Lora]

Could anyone explain to me how to solve this issue?

I have run a regression and stored it but I cannot use it to perform tests. Here is the relevant code: the fist code line shows that the equation does exist and has been restored, but the second code line shows an error as though the equation were not there. (I know that I could write simply "test laidoff_event", but I need to perform a test of coefficients that come from different regressions). Thank you

. est restore reg2_y3_fem
(results reg2_y3_fem are active now)

. test [reg2_y3_fem]laidoff_event
equation reg2_y3_fem not found
r(303);

[William Lisowski]

I think you misunderstand the difference between the "equation" and the "estimates".

Some modeling commands will estimate more than one equation from a given command. The stored estimates from those commands include equation names created by the command.

The name you provide in

estimates store reg2_y3_fem

is the name of the stored results (the estimates) not an equation name.

Note that the response

. est restore reg2_y3_fem
(results reg2_y3_fem are active now)

tells you that the reg2_y3_fem results were restored - not that a reg2_y3_fem equation was restored.

I know that I could write simply "test laidoff_event", but I need to perform a test of coefficients that come from different regressions

Only the most recent set of estimates - from the most recent estimation or estimates restore command - will be in memory at a given time, so you do not need to distinguish between coefficients from different regressions - only one set of coefficients is available.

[Joro Kolev]

Another misunderstanding here is that if you store estimates you can after that perform on them whatever tests you want.

To perform joint test, you need to use something like -suest-, some command that allows joint estimation.

[Eduardo Lora]

Thanks a lot, I think I got it now.