Page 1 of 1

Solving consistently with known endogenous variables

Posted: Fri Mar 10, 2017 3:39 am
by ferdy_adam
In forecasting, a practical question often arises:
Let’s presume t is the first period for which the model is solved i.e. all endogenous variables are unknown.
Then, if we have information on endogenous for t or t+x, we can incorporate them in the solution with the model.exclude(actexist=t) statement.
If however we have the following situation:
x = y + z
and if we observe x (which is an identity) but not y and/or z (which are endogenous behavioral), how can we guarantee the holding of the identity? This is a very practical question since in Macroeconomics, we often observe the aggregates before we observe the detailed components.
Many thanks