Calibrating measurement errors

I don’t understand how to calibrate measurement errors in Dynare. I can think of possible ways around this (really tight prior), but I just wanted to check if there is a "straight " way to calibrate measurement errors. Computationally this would be superior.


There is currently no facility in Dynare to calibrate measurement errors. But measurement errors can always be treated as structural shocks, then it is a question of interpretation of the results.

Add measurement errors to the observed variables, either directly in the model if possible or by definining additional variable for the observable variables. Then specify their standard deviation via the shocks block



Hi Michel

Currently there is the M_.H object in Dynare that is suppose to store the calibrated standard deviations of the measurement errors. How I can declare those standar deviations in Dynare? It seems that object is important for running the kalman filter with smoothing errors (DsgeSmoother).