Hi Everyone,
When Dynare estimates some parameters using Bayesian methods, it does fix the other parameters even if some of the other parameters were calculated based on the initial values of the estimated parameters.
The problem is that Dynare does not update these related parameters automatically during the estimation process, unless you replace them by explicit functions of the estimated parameters. However, when a model is large with many parameters related to the estimated ones, it is hard to to do so. How to solve the problem?
Thanks a lot in advance,
Bob
Take a look at Pfeifer(2013): “A Guide to Specifying Observation Equations for the Estimation of DSGE Models” sites.google.com/site/pfeiferecon/Pfeifer_2013_Observation_Equations.pdf
Short answer: either use model-local variables or a steady state file to update/take care of the parameter dependencies. There are no other ways.
Johannes,
What I meant is that: do not define dependent parameters, but rather write them explicitly in terms of estimated parameters. That is definitely a way!
For example instead of defining a parameter Rbar=1/beta, write 1/beta directly.
However, using model-local variables helps write a model more neatly. That’s great! Thanks a lot.
Bob