Using Stata 13's -forecast- with supply and demand models

[Trevor Zink]

Apologies if this is exceedingly dense of me. I'm trying to use -forecast- to make a supply and demand model the individual equations of which I have already estimated using -ivregress-

As I understand it, the procedure for making models involves linking saved equation estimates, the left-hand sides of which are treated as endogenous. Identities can be added with -forecast identity-
Supply and demand models present two complications:
1) Along with supply and demand, price is also endogenous, and it appears on the right-hand side of the estimated equations. Specifying that price is endogenous in the -ivregress- or -reg3- estimation is ignored by -forecast-. How do you tell -forecast- that price is endogenous and needs to be solved for?
2) The market-clearing identity that supply = demand is required. However, it is not possible to input this identity to -forecast- because it does not permit variables to be added on the left-hand side twice.

I'll illustrate with an example:

. forecast create example, replace
  (Forecast model al1 ended.)
  Forecast model example started.

. forecast estimates EQps
  Added estimation results from ivregress.
  Forecast model example now contains 1 endogenous variable.

. forecast estimates EQpd
  Added estimation results from ivregress.
  Forecast model example now contains 2 endogenous variables.

. forecast describe endogenous

Forecast model example contains 2 endogenous variables:
-----------------------------------------------------------------
       Variable      |       Source        |    # adjustments
-----------------------------------------------------------------
  1. l_qsprim               estimates                  0
  2. l_qdprim               estimates                  0
-----------------------------------------------------------------

. forecast identity l_qsprim = l_qdprim
cannot add identity
    Endogenous variable l_qsprim has already been added to the model. You cannot
    add an endogenous variable to the model multiple times. Type forecast describe
    endogenous to see how the variable was added to the model.
r(110);

[Trevor Zink]

Bump for visibility. I'd really appreciate any insights here!

[Trevor Zink]

Anyone? Please?

[Nick Cox]

I can't say why, but people who use this kind of command don't seem to be at all active on Statalist. I have never used it, which is personal reason enough not to answer.

On a different level, we don't have rules on Statalist as a web forum now about bumping, but we used to have explicit advice about re-posting, and the advice was one re-post only please. It's absolutely not a criticism that you are not aware of advice that doesn't exist, but if you don't get an answer on this bump, sorry, but I doubt you'll get anywhere.

It is true that we ask that people use full real names on Statalist; you should have read that advice in the FAQ.