Browse Prior Art Database

Programming System to Solve Decisional Linear Models

IP.com Disclosure Number: IPCOM000075298D
Original Publication Date: 1971-Aug-01
Included in the Prior Art Database: 2005-Feb-24
Document File: 3 page(s) / 16K

Publishing Venue

IBM

Related People

Stajano, A: AUTHOR

Abstract

This system, subsequently referred to as DMS (Decisional Models Solution), is a programming system planned as a tool for the economic research worker in testing a model and studying, by its means, the economic system's behavior.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 47% of the total text.

Page 1 of 3

Programming System to Solve Decisional Linear Models

This system, subsequently referred to as DMS (Decisional Models Solution), is a programming system planned as a tool for the economic research worker in testing a model and studying, by its means, the economic system's behavior.

DMS has been designed in order to solve a multiequational linear model with the following structure:

(Image Omitted)

Only nonzero matrix elements need be inputted; one or two matrices out of A, B, B may be absent; e.g. in the case of the Input/Output models only the matrix B is present.

Matrix coefficients are stored by columns, regardless of whether they belong to A, A(p), B or B(p) which is specified by control cards at solution time. This enables the usage of the same data in different economic analyses:
1) analyses in which distinct and isolable subsets of the

models are solved or in which one or several rows and/or

columns are suppressed.
2) analyses in which the role of instrument and endogenous

variables is interchanged

to determine which value should be assigned to one

or several instrument variables to achieve a given target.

The economic research worker will determine whether this change should require an off-line re-estimation of the model. The research under point 2) above can be developed by the DMS user in various ways. The analyst fixes his economic objective by assigning target values to one or several endogenous variables. He also chooses which instrument variables should be used to achieve his aims.

The number K of instruments will usually be not less than the number M of the targets. If K = M the analysis is developed as follows:
1) The instrument variables are defined as endogenous, i.e.

dependent
2) The target variables are defined as exogenous, i.e.

predeterminate
3) The target variables are given the required values in the

time series file
4) The solution is requested.

If on the contrary the number K of instruments is greater than the number M of targets two distinct solutions are possible:
a) with a procedure similar to the case above, all possible

groups of M instrument

variables are selected and defined as unknown while the

remaining K-M instrument variables are given the values

1

Page 2 of 3

previously inputted in the time series file.

The system permits a reduction of the number of

selections (whose number

(/K/(M)) is remarkable also for small values of K and K-M),

in fact the economist

will be often able to recognize as meaningless

some selections requiring the simultaneous intervention on

given couples

of instruments. According to the statement of such

restrictions, the system reduces the number of generated sets

of procedures.

The comparison among the various solutions may turn out to

be of great interest

to evaluate the financial and/or political cost of

the achievement of the stated targets. But sometimes this

comparison may become

too detailed at least as a first approach. In such a

case it may be advantageous to resort...