The Closure report file

The Closure report file consists of several different sections:

1. Directly matching equations and variables

This consists of a list of variables and equations which have matching names. If the dimensions do not match, TABmate follows the pair with the the message Size !?. This does not always indicate a problem (maybe two different sets have the same size). The variables will be endogenous in the closure that TABmate constructs.

2. Matching equations and variables (many equations to 1 variable)

This consists of groups of equations which together explain a single variable. Its your job to check that the total size of the several equations adds up to to the size of the variable they explain. The variables will be endogenous in the closure that TABmate constructs.

3. Unmatched equations

This consists of a list of equations which TABmate could not link to any variable. The fewer of these, the better (see paragraphs 4 and 7 below).

4. Unmatched variables

This consists of a list of variables which have no equation named after them. These variables will be exogenous in the closure that TABmate constructs. The list is formatted so that you can paste it into a CMF file.

If you have some unmatched equations (see above) the list of unmatched variables will be too long to use as an exogenous list: it will include endogenous variables of the same total size as the unmatched equations. You will have to manually remove these variables from the list of unmatched variables. Better, rename the unmatched equations so that TABmate does this for you automatically.

Even if the exogenous list suggested by TABmate is of the correct size, it may not be quite the closure that you want to use for some particular simulation. You may wish to swap exogenous and endogenous variables. However it is normally easier to modify an existing closure than to create a new one from scratch.

5. Backsolved variables

This consists of a list of backsolved variables which GEMPACK has already associated with particular equations. Backsolved variables are always endogenous.

6. Condensation suggestions

This consists of all those variables from Section 1 (directly matching equations and variables) which have array dimensions. TABmate suggests that you backsolve them, but often substitution may be more appropriate. You probably would not want to backsolve or substitute all the variables which TABmate lists -- but you should be aware that very large variables, if not substituted, can increase solution times and memory requirements. On the other hand, substituted variables will be invisible in both the solution file and AnalyseGE.

A better condensation can make large models run more quickly. The Closure Report can be used with the Condensation Information File to help you find such a condensation.

7. Tally of variables and equations

If TABmate can link every equation to some variable (that is, if there are no unmatched equations in section 3 above) a table is produced showing how many equations and variables (either exogenous or endogenous) there are of each size. This is often useful for documenting the model.

In addition, if set sizes are known, the table will tell you how many individual variables and equations the model contains. Use these numbers to impress people at parties !

8. Strangely named variables

FInally there is a list of strangely named variables: variables that are ordinary change although this is not suggested by their name; or are percent change although their name suggests otherwise. The aim is to help you notice a common error: forgetting to add the (change) qualifier to a variable.