On the plausibility of using linear programming to trace important input-output coefficients in the framework of tolerable limits

Recent input-output (IO) literature offers original proposals on using linear programming (LP) to make 'tolerable limits' approach suitable for measuring the importance of IO coefficients to an economy. In this paper, I focus on one of such influential proposals presented in Tarancon et al. [(2008). A revision of the tolerable limits approach: searching for the important coefficients. Economic Systems Research, 20, 75-95]. In the theoretical part of this paper, I provide exact analytical solutions to the LP problems formulated in Tarancon et al. The main result proves that the classification of IO coefficients with respect to their importance in the sense of the LP-based indicators of Tarancon et al. does not depend on the benchmark welfare measure of interest. This fact, in turn, severely reduces practical applicability of the discussed LP-based approach to tracing important IO coefficients.