Abstract
For the maintenance and development of a disaggregated econometric model, the present paper deals with some aspects related to the identification of important intermediate transactions. First, some methodological aspects related to earlier works are discussed. By comparing the properties of different selection methods based on the recent West German input–output table, the calculations confirm the common finding that the necessary level of endogenization can be bounded meaningfully by an a priori investiga tion of the intersectoral transactions. This is also valid if the selection method considers restrictions that deal with the consistence of input–output tables and if some broader error criteria are applied. The main result of the empirical applications seems to be that some attention should be paid to temporal changes of important coefficients.