ADAPTATION OF DETAILED INPUT-OUTPUT INFORMATION: RESTRUCTURING AND AGGREGATION

This article deals with problems of construction of "square" input-output tables from detailed commodity and industry data and operationalization of the tables for use in econometric modeling. The adaptation procedure suggested is quite new and involves perfect and imperfect aggregation, and suppression of insignificant cells of the tables.
Using slight modifications of well-known input-output methods (to make definitions conform to general concepts of network flow theory) it is shown that the construction and aggregation of tables, as well as the suppression of minor cells, can be viewed as still higher levels of the very same process: the search for a manageable model with roughly the same abstract properties as the original detailed, but overwhelmingly large, model.
Simultaneously a consistent input-output terminology is suggested with fewer symbols and more rules than usual.
The adaptation procedure has been applied successfully to the 1982 version of ADAM, the macroeconometric model operated by Danmarks Statistik (the Danish Central Bureau of Statistics).