Input-output models are mathematical models which examine the relations
within an economy in numerical terms. The fundamental part of this model is
the input-output tables displaying the flow of goods and services. These inputoutput
tables include the statistics of sector production, gross domestic product,
foreign trade and sector value added, complied through different statistical
resources. All these data are summed in input-output tables within a frame of
consistency. With the input-output tables, it is possible to analyze and perform
manufacturing planning, price and cost, sector investment and productivity as
well as structural analyses of economy and making comparisons among the
terms. Besides, the revision in national product calculations is available
through the input-output tables. Despite all these, it is a costly and timeconsuming
process to prepare input-output tables. Hence, it is not always
possible to yearly-prepare input-output tables. Because of this, the use of inputoutput
tables in model and economic planning leads to delays and outdated
information.
The input coefficients, also known as technology coefficients, calculated in
input-output (I-O) tables also suffer from the delays and outdated information
and data. For this reason, some methods were developed in preparing the inputoutput
tables to get rid of the problems mentioned above. One of the methods
employed in updating the input coefficients is linear programming model. In the
estimation of input-output models with linear programming methods, there are
different methods used by different researches. In this study, created through
linear programming method, goal programming model was employed. Another
method commonly employed in the estimation of input coefficients is RAS. This
method in the simplest terms is the re-compilation of the latest input-output
table in manner that will yield the sum of input and output for the targeted year
The main objective of this study is to obtain the technology coefficients
published in 2002 based on the 1998 table of input-output within the framework
of mathematical modeling such as RAS and linear programming.It is observed that there is no standard in the input-output tables published
in Turkey. The only table that complies with the European standards is the table
published in 2002. In order for the mathematical models to be implemented, the
table of 1998 should be in this standard too and in the same sector number.
Hence, firstly using the conversion key prepared by TURKSTAT, the inputoutput
table of 1988 was upgraded to the standards of European Union.
In the analysis of the models, the table of 1988 input-output was considered
as the base table, and the table of 2002, on the other hand, was considered
target year input-output. The ij a values, denoting technology coefficients, were
calculated through the table of 1998 and column and row sums related to sector
intermediate input were computed using the input-output tables of 2002. For the
estimation of the technology coefficients RAS and goal programming models
were set up and the analyses were performed through LINGO 8 software. For
RAS analysis, package program used by TURKSTAT was made use of.
In order for the estimation result to test their achievement level, the Theil’s
inequality test coefficient was used, which shows the cohesion level between the
real values and estimated values under consideration. This inequality coefficient
was computed along with Theil inequality coefficient subcomponents in order to
examine the extent to which the estimated input-output coefficients capture the
real coefficients. A value near to 0 (zero) shows the success of the estimation,
while a near 1 value shows the opposite.
When the achievement test results of the models used are examined, the most
successful model is found out to be RAS model in computing the technology
coefficient of the year 2002. Then follows goal programming.
