Social Accounting Matrix Balanced Based on Mathematical Optimization Method and General Algebraic Modeling System
Ming- Chang Lee *
National Kaohsiung University of Applied Sciences, Taiwan, No. 415 Chien Kung Rd. Kaohsiung Taiwan
Li- Er Su
Shih Chien University, Kaohsiung Campus, Kaohsiung Taiwan, 200 University Road, Neimen, Kaohsiung Taiwan
*Author to whom correspondence should be addressed.
Abstract
Aims: Social accounting matrix (SAM) is a kind of important form in the national accounts calculation. The requisite data are typically collected from a number of different sources. Therefore, much of data used will not be immediately consistent. The aim of this paper present the various methods of balancing applied to social accounting matrix.
Design: These methods (RAS, Cross-entropy (CE), least square, linear programming with and linear programming with method) are applied widely for develop and keep and time consistent multi-sector datasets.
Methodology: This paper solves the social accounting matrix balanced problem using iterative method (in RAS approach) and using the computer program about General Algebraic Modeling system (GAMS) (in other approach). The sum square error of the estimated table from the initial table is used as the main instrument to measure deviation of the updated matrices from the matrix.
Results: We found that CE method is more accurate than the other methods.
Keywords: Social accounting matrix (SAM), System of National Accounts (SNA), Cross entropy (CE), least square method, Linear programming norm L1, linear programming normL∝