Strange Error Message

[Daniel Bendel]

Hey all,

I got the following error message when using model(block):

Configuring Dynare ...
[mex] Generalized QZ.
[mex] Sylvester equation solution.
[mex] Kronecker products.
[mex] Sparse kronecker products.
[mex] Local state space iteration (second order).
[mex] Bytecode evaluation.
[mex] k-order perturbation solver.
[mex] k-order solution simulation.
[mex] Quasi Monte-Carlo sequence (Sobol).
[mex] Markov Switching SBVAR.

Starting Dynare (version 4.4.2).
Starting preprocessing of the model file ...
Found 39 equation(s).
Evaluating expressions...done
Computing static model derivatives:
- order 1
Normalizing the model...
Finding the optimal block decomposition of the model ...
7 block(s) found:
6 recursive block(s) and 1 simultaneous block(s).
the largest simultaneous block has 25 equation(s)
and 25 feedback variable(s).
Computing dynamic model derivatives:
- order 1
- order 2
Normalizing the model...
Normalization failed with cutoff, trying symbolic normalization...

This application has requested the Runtime to terminate it in an unusual way.
Please contact the application's support team for more information.
terminate called after throwing an instance of 'DataTree::UnknownDerivIDException'

What does that mean??

[jpfeifer]

That's a Dynare bug. Please send me the model file.

[Daniel Bendel]

I have sent the files. Many thanks in advance!!

[jpfeifer]

Thanks, we are going to look into it. For the moment, please delete the (block) statement and fix

MODEL_DIAGNOSTICS: The following endogenous variables aren't present at the current period in the model:
Rdb

[Daniel Bendel]

okay.

[jpfeifer]

When fixing that Rdb does not appear at the current period, the block option should work. Future versions of Dynare will provide a more explicit error message.

[Daniel Bendel]

Why is it a problem for dynare that a variable does not appear at the current period? Can one solve it with the EXPECTATION operator of dynare?

[jpfeifer]

It is not a Dynare problem, but a problem with the solution concept in general. The goal is to solve the equilibrium system to determine the value of time t variables as a function of the states. If a variable at time t does not appear, the solution for this variable cannot be determined as it does not play a role at all. The expectations operator is no solution. You want to get the value of the variable and not of its expectations. If you are actually only interested in the expectatation, get rid of the underlying variable and define the expectation as a separate variable.

[Daniel Bendel]

okay I understand the first thing you've said. But what do you mean with "If you are actually only interested in the expectatation, get rid of the underlying variable and define the expectation as a separate variable."? Say I have a variable Y, and Y is only persent in t+1, so how should I get rid of it and redefine it?