A method for estimating the sampling variances of multipliers derived from a from-to model

Abstract

Interindustry models provide a) a simple, logical, and internally consistent way of portraying the interdependent structure of a region and b) an analytic framework which can be used to provide answers to a wide variety of questions. Many of these answers can be expressed as easily understood multipliers, but because the persons for whom these multipliers are most useful possess only limited understanding of the underlying model, they are often forced to accept (or reject) them on faith. In many cases it should be possible to supply these users with some indication of the reliability of estimated multipliers. Many factors affect reliability and these factors are not equally measurable. This paper has the limited objective of estimating that component of multiplier variance induced by the use of sample data to construct the basic model. The procedure adopted exploits a first order approximation technique. It is applied to the data used to construct a from-to model of Oklahoma Planning Region 9. Selected limitations and implications of the resulting estimates are examined.