An Interregional Perspective on Aggregation Bias and Information Loss in Input-output Analysis

This paper extends previous theorems concerning aggregation bias in single-region and national input-output models to interregional input-output (IRIO) models. The results indicate that the conditions for zero aggregation bias are more stringent in input-output models involving multiple regions. In addition, the paper derives the relationship between the information content of IRIO and national input-output accounts. The results confirm the conventional wisdom that the information content of multiple region input-output models of the U.S. economy contain more information than national input-output models.