Abstract
The primary aim of this paper is to extend my previous work on the reconciliation of regional coefficients in input-output models. Specifically, an effort is made to show how certain types of a priori or judgmental information may be incorporated into the minimum variance reconciliation (MVR) process. In addition, this paper provides a response to Miernyk's objections that the MVR method: (1) produces coefficient estimates of doubtful statistical or theoretical validity and (2) systematically underestimates diagonal regional coefficients as well as the sum of interindustry sales for a given sector.

This publication has 6 references indexed in Scilit: