Same Closure for One Year of Rerun and Policy


[This is a technical topic. Only experts need read it.]

Note that RunDynam only produces one .BOT file (and runs only one bot job) for each year of the Policy.

If you are rerunning the Base with Policy closures, RunDynam uses the same .BOT file used for any year of the Rerun and for that same year of the Policy. This is because RunDynam assumes that the closure for any year of the Policy is the same as the closure for the same year of the Rerun. If you have data-dependent sets in your model, you may violate this assumption in certain rare cases, as explained below.

Suppose, for example, that the set TRADEXP of traditional export commodities in your model depends on what fraction of total production of each commodity is exported. Then the commodities in this set may change depending on the starting data for the simulation. Imagine an application in which commodity C10 was not a traditional export commodity in the initial data, and suppose that it remains a non-traditional export throughout the Base (and hence the Rerun). Now suppose that the Policy shocks are so favourable to exports of this commodity C10 that it gets into the set TRADEXP after 8 years of the Policy run. Then, if the closure depends on this set TRADEXP, the closure for the ninth year of the Rerun and for the ninth year of the Policy would be different.

The results you get via RunDynam for the ninth year of the Policy would be suspect since RunDynam uses the same .BOT file for this ninth year of the Rerun and of the Policy.

At present RunDynam does not check that the closures are the same for any year of the Rerun and the same year of the Policy. It is your responsibility to ensure that.

The effects of the policy are the differences between each year of the Policy and the corresponding year of the Rerun. The whole rationale for doing the Rerun is to remove accuracy changes and path-dependency changes from these differences. If you have a different closure in these two runs, you are negating the reason for doing the Rerun.



URL of this topic: www.copsmodels.com/webhelp/rundynam/hc_sameclos.htm

Link to full GEMPACK Manual

Link to GEMPACK homepage