Abstract

A Computational Routine for Disaggregating Industry Margin Data to Estimate Product Margin Rates by Matthew D. Atkinson
To be presented at the Fall 2003 Federal Committee on Statistical methodology Research Conference, Washington, D.C.

Retail industry product margin rates are used to estimate the retail output proportion of final consumption commodities. The Census Bureau collects data on industry margin rates, but it does not collect product margin rate data. To estimate retail industry-by-commodity output, industry margin rates are disaggregated by product. A number of controls are available for disaggregating industry data. This paper introduces a formal computational method for disaggregating industry margin data using Bayesian statistics and simulation. The routine is capable of accurately imposing multiple controls simultaneously. The method's accuracy is demonstrated by an evaluation of its industry product margin rate estimates. In addition to producing accurate disaggregate estimates, the method is fast and its estimates are replicable. The computational method has a broad range of applications beyond the estimation of industry-by-product margin rates.

Last changed: April 14, 2003