Hi All,
I am working with a (confidential) dataset that conains in it an explanatory variable, two covariates, and a bunch of dummy variables, corresponding to year and country fixed effects. I wish to estimate this model using Robinson's root n consistent estimator (partially linear model). In two separate commands, I have generated dummy variables for the country and year effects using for instance:
tab year, gen(y)
and similarly for country:
tab country, gen(i)
I then use the command:
semipar y x1 i* y*, nonpar(x2)
where the covariate x2 is allowed to be non parametrically identified.
However, stata returns the error message:
__00000I not found
This is quite perplexing as when I estimate the model linearly, the commands work just fine. Is this a problem specific to the semipar command?
I am working with a (confidential) dataset that conains in it an explanatory variable, two covariates, and a bunch of dummy variables, corresponding to year and country fixed effects. I wish to estimate this model using Robinson's root n consistent estimator (partially linear model). In two separate commands, I have generated dummy variables for the country and year effects using for instance:
tab year, gen(y)
and similarly for country:
tab country, gen(i)
I then use the command:
semipar y x1 i* y*, nonpar(x2)
where the covariate x2 is allowed to be non parametrically identified.
However, stata returns the error message:
__00000I not found
This is quite perplexing as when I estimate the model linearly, the commands work just fine. Is this a problem specific to the semipar command?
