Condition Plan

[pfjulio]

Dear Jaromir,

I used to run this piece of code in earlier versions of IRIS, but now I am getting an error. Has the syntax changed or am I doing anything wrong?

pshadow = Plan(m,qq(2022,1):qq(2025,4),'anticipate=',false);
pshadow = condition(pshadow,qq(2022,1):qq(2025,4),'GDP_GRWTH');

I get the following error:

*** This name cannot be conditioned upon in the simulation Plan: GDP_GRWTH

This error holds for any endogenous variable I select.

Best
Paulo Júlio

[jaromir-benes]

Could you please post the files? Jaromir

…

On Mon, Nov 14, 2022 at 4:42 PM pfjulio ***@***.***> wrote: Dear Jaromir, I used to run this piece of code in earlier versions of IRIS, but now I am getting an error. Has the syntax changed or am I doing anything wrong? pshadow = Plan(m,qq(2022,1):qq(2025,4),'anticipate=',false); pshadow = condition(pshadow,qq(2022,1):qq(2025,4),'GDP_GRWTH'); I get the following error: *** This name cannot be conditioned upon in the simulation Plan: GDP_GRWTH This error holds for any endogenous variable I select. Best Paulo Júlio — Reply to this email directly, view it on GitHub <#351>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AGCVKKVE2HZWWZMRQ262PSTWIJMUXANCNFSM6AAAAAAR76VLCU> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[pfjulio]

Code_model.zip

Here is a simplified version of the code where this occurs.

[jaromir-benes]

Hi - the syntax has indeed changed (sorry about that - but for a good reason :) p = Plan.forModel(m, range, method="condition"); p = exogenize(p, qq(2020,1):qq(2021,4), "variable_name"); p = endogenize(p, qq(2020,1):qq(2022,4), ["shock1_name", "shock2_name", ...]); How it works: Conditioning means simply exogenizing a smaller number of variables than the number of shocks. This is by default not allowed in Iris (because this is rather a rare case, and from my practical experience, if you have #variables<#shocks, it's most of the time a mistake by the coder/economist, so Iris throws an error for your benefit). To allow conditioning, you have to specify that explicitly when creating the Plan object, hence method="condition" (by default, method="exogenize" meaning you have to have #variables==#shocks) Then, you exogenize/endogenize whatever you want. Only those shocks explicitly listed will be used. This last point was exactly the reason why we abandoned the function "condition" - it didin't allow to easily select the shocks participating in the conditioning (you had to do it through their stds). Lastly, note that the date ranges can be different for each variable and each shock... You could e.g. have p = Plan.forModel(m, range, method="condition"); p = exogenize(p, qq(2020,1):qq(2021,4), "variable_name"); p = endogenize(p, qq(2020,1):qq(2022,4), ["shock1_name", ...]); p = endogenize(p, qq(2022,3):qq(2023,1), ["shock2_name", ...]); And of course, the anticipation status of the shocks matter. Hope this helps Best Jaromir

…

On Mon, Nov 14, 2022 at 6:27 PM pfjulio ***@***.***> wrote: Code_model.zip <https://github.com/IRIS-Solutions-Team/IRIS-Toolbox/files/10005499/Code_model.zip> Here is a simplified version of the code where this occurs. — Reply to this email directly, view it on GitHub <#351 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AGCVKKUQBO4BAX77NVY2YT3WIJY67ANCNFSM6AAAAAAR76VLCU> . You are receiving this because you commented.Message ID: <IRIS-Solutions-Team/IRIS-Toolbox/repo-discussions/351/comments/4139183@ github.com>