The moments of simulated variables in Dynare for a variable is based on population or sample?
We know that standard deviation and variance of sample and population are different. In MATLAB I calculated these moments for a typical endogenous variable, for sample and population but the results were different from dynare output result.
If you use periods=0
(the default), then populations moments are used. Otherwise, it is sample moments.
I set periods=50000 and irf=80 then I tested the sample standard deviation and variance of simulated variables in Dynare for example Y, C,
I, L,… but the results were not the same.I used MATLAB bulit-in functions for calculation of the moments, for example std(Y,0) for standard deviation of Y for sample but it’s result was not the same with Dynare output.
What exactly did you do? Did you use a filter? And how different are the results?
I simulated my model for periods=50000 and irf=80. Dynare gives output for example standard deviations of simulated variables in the model. For example Y,C,I ,L,W,…
I calculated standard deviation of Y with MATLAB built-in function for Y. I calculated standard deviation for population and sample, but the result was not the same with Dynare standard deviation for Y.
The difference is not significant at all. But I don’t know how Dynare calculates the simulated variables moments.
How do you compute the population moments from a simulated series? Also note that Dynare by default drops the first 100 observations when computing moments. See the manual on the drop
-option
I did not know this point.Thank you so much professor.I think that my problem solved because I did not know this important point.