Hi

I am working on a CGE model. I received this message

********************************************************************************

**** ERROR: CNS models must be square

This model contains 3270 equality rows and 3274 unfixed columns

MODEL STATISTICS

BLOCKS OF EQUATIONS 105 SINGLE EQUATIONS 3,270

BLOCKS OF VARIABLES 96 SINGLE VARIABLES 3,274

NON ZERO ELEMENTS 15,180 NON LINEAR N-Z 7,319

DERIVATIVE POOL 20 CONSTANT POOL 1,367

CODE LENGTH 20,968

SINGLE =E|X= EQUS 3,270 SINGLE UNFIXED VARS 3,274

*********************************************************************

How I could find the extra variables and fix the problem?

## fixing a non square system in CNS

### Re: fixing a non square system in CNS

Hi

If you have your equations cast in the typical Arrow-Debreu-Format with zero-profit conditions, market clearing conditions and income definitions, the only thing you can do is to map your variables to these equations (e.g. respectively production level, prices and incomes) and see if the number of (non-fixed) mapped variables equals the number of equations. Note, that you might have auxiliary equations for defining variables used in other equations (like L(i) =E= .....) for labor demand in a sector to be used in the market clearing equation.

Cheers

Renger

If you have your equations cast in the typical Arrow-Debreu-Format with zero-profit conditions, market clearing conditions and income definitions, the only thing you can do is to map your variables to these equations (e.g. respectively production level, prices and incomes) and see if the number of (non-fixed) mapped variables equals the number of equations. Note, that you might have auxiliary equations for defining variables used in other equations (like L(i) =E= .....) for labor demand in a sector to be used in the market clearing equation.

Cheers

Renger