Evaluating the Relative Performances and Comparing the Forecasting of Thailand's Technical Coefficients Covering 2016-2025 by MTT and RAS

Abstract

Dealing with the timeliness of Thailand's technical coefficient tables, this research contributes the methodology by forecasting them by Matrix Transformation Technique (MTT) and RAS, determining the relative accuracy of these two methods. Regarding MTT, rather than predicting all elements of the transformed matrix as originally purposes, this research forecasts only some significant elements which have evidence of change generating a high correlation to the technical coefficients. For the evaluation, the analysis first compares the updating performance in relation to 2015, considering the cell-by-cell (or partitive) and whole matrix (or holistic) levels. The result shows that MTT is outstanding for both sector and whole matrix levels; however, when considered cell-by-cell, RAS is superior over MTT. Hereafter, the study uses MTT and RAS to forecast Thailand's technical coefficient tables covering the period 2016-2025. Most elements (71.50%) were predicted in the same direction. However, the remaining 28.50% showed a difference in the direction of the forecast. Hence, the cell-by-cell assessment for these 73 elements was also checked, finding that MTT offers better performance for 38 elements, while RAS exhibits better performance for 35 elements.