Prior distribution for parameter rhoA has unbounded density!

Estimation ML/Bayesian estimation
1

Dear Professor,
I am working on a midscale DSGE, and the thing is, calibrated model is working (with clear IRFs, except occurrence of a unit root, with Dynare version 5.3）， but when i run the estimated model, errors break out:
’
Prior distribution for parameter rhoA has unbounded density!
Prior distribution for parameter rhoG has unbounded density!
Prior distribution for parameter rhoB has unbounded density!
Prior distribution for parameter rhox has unbounded density!
Prior distribution for parameter rhof has unbounded density!
Prior distribution for parameter rhob has unbounded density!

DYNARE_SOLVE (solve_algo=2|4|12): the Dulmage-Mendelsohn decomposition returned a non-square block. This means that the Jacobian is singular. You may want to try another value for solve_algo.
Error in computing likelihood for initial parameter values

ESTIMATION_CHECKS: There was an error in computing the likelihood for initial parameter values.
ESTIMATION_CHECKS: If this is not a problem with the setting of options (check the error message below),
ESTIMATION_CHECKS: you should try using the calibrated version of the model as starting values. To do
ESTIMATION_CHECKS: this, add an empty estimated_params_init-block with use_calibration option immediately before the estimation
ESTIMATION_CHECKS: command (and after the estimated_params-block so that it does not get overwritten):
’
May i ask, do you have any clue that what should i do next to debug or fix this?
Yours sincerely.

P.S. the files:
Uploading: LRtest_swaddxi.mod…
Uploading: LR_sw_xim_esti.mod…
mydata.m (7.1 KB)(data for estimation, raw macro data have been processed for growth rates and logarithms.)

2

LRtest_swaddxi.zip (3.9 KB)(calibrated)
LR_sw_xim_esti.zip (4.3 KB)(estimated)
(the mod file cannot be uploaded directly, so i re-upload their zip format).

3

That message signals a problem in steady state finding for the initial parameter draw. It seems just providing initial values is not sufficient, but rather you need to provide an analytical steady state.

5

I appreciate your reply, and may i ask, does manual calculation of steady state imply the “analytical steady state”?(Cause i don’t understand what is an analytical steady state :slight_smile: sry) Just like the figure below:
微信图片_20230820160755(‘var+s’ stands for steady state of variables)

6

Yes, having explicit values is the difference from a numerically computed steady state. But it’s important to update these values for each parameter draw, i.e. it needs to be in a steady_state_model-block.

7

Thanks professor. After refering to your hard-working about specifying Observation Equations and dynare graphs, my files finally work (by adjusting observation equations and changing priors).
Yours sincerely.

1 Like