!log_variables
VariableName, VariableName,
VariableName, ...
!log_variables
!all_but
VariableName, VariableName,
VariableName, ...
!log_variables
VariableName, VariableName,
VariableName, ...
<REGEXP>, <REGEXP>, ...
List all log variables under this headings. Only measurement or transition variables can be declared as log variables.
In non-linear models, all variables are linearised around the steady state or a balanced-growth path. If you wish to log-linearise some of them instead, put them on a !log_variables
list. You can also use the !all_but
keyword to indicate an inverse list: all variables will be log-linearised except those listed.
To create the list of log variables, you can also use regular expressions, each enlosed in a pair of angle brackets, <
and >
. All measurement and transition variables whose names match one of the regular expressions will be declared as log variables. See also help on regular expressions in the Matlab documentation.
The following block of code will cause the variables Y
, C
, I
, and K
to be declared as log variables, and hence log-linearised in the model solution, while r
and pie
will be linearised:
!transition_variables
Y, C, I, K, r, pie
!log_variables
Y, C, I, K
You can do the same job by writing
!transition_variables
Y, C, I, K, r, pie
!log_variables
!all_but
r, pie
We again achieve the same result as above, but now using a regular expression.
!transition_variables
Y, C, I, K, r, pie
!log_variables
<[A-Z]\w*>
The regular expression [A-Z]\w*
selects all variables whose names start with an upper-case letter. Hence, again the variables Y
, C
, I
, and K
will be declared as log variables.