Page 1 of 1

A question about measurement errors

PostPosted: Thu Apr 29, 2010 7:36 pm
by drodrigu
Hi, I like to know if there is a vay to fix de measurement errors in a estimation without construct equations like this one?

Y_obs = y_t + measurement_error,

thanks,

drodrigu

Re: A question about measurement errors

PostPosted: Fri Apr 30, 2010 6:05 am
by jpfeifer
Hi drodrigu,

what do you mean with "fix the measurement error"?

Re: A question about measurement errors

PostPosted: Fri Apr 30, 2010 3:02 pm
by drodrigu
What I am triying to do is to make an estimation of the structural parameters given the variance of the measurament errors. That is, I dont want to estimate the variance of the measurament errors.

Re: A question about measurement errors

PostPosted: Fri Apr 30, 2010 7:02 pm
by jpfeifer
Hi,

that's possible, but you still have to specify a measurement equation like the one you wrote above. You then treat the standard deviation of the measurement errors like any other non-estimated parameters. You specify their standard deviations as for all exogenous shocks in the

Code: Select all
shocks;
end;


block. When you then omit the measurement errors in the

Code: Select all
estimated_params;
end;


block Dynare will treat them/their standard deviation as given and will not estimate them.

Re: A question about measurement errors

PostPosted: Fri Apr 30, 2010 8:16 pm
by drodrigu
Hi jpfeifer

Thanks for your response

Thanks

drodrigu