SFB 303 Discussion Paper No. A - 346

Author: Gyarfas, Gabor, and Hermann Quinke
Title: Ein Verfahren zur Konstruktion synthetischer Mikrodaten aus aggregierten Stukturinformationen
Abstract: Simulation analyses of the tax-transfer system require disaggregated data. If microdata are not available - e.g. because of privacy issues - often a large number of published tables (aggregate structural information) can be utilized. In addition to these "hard" data there are frequently more or less "soft" prior notions based on other (older or foreign) data sources or simply on ad hoc reasoning. Our paper presents a method for the construction of synthetic microdata from aggregate structural information and a prior distribution. Such a database is not only consistent with the available information, but it also has desirable statistical properties in the case of large samples. We prove that the empirical distributions corresponding to the synthetic samples converge to the distribution that is consistent with the aggregate information and maximizes the entropy relative to the prior distribution. Two small Monte Carlo experiments illustrate this proposition.
Creation-Date: October, 1991
Unfortunately this paper is not available. Please order a hardcopy via e-mail.

SFB 303 Homepage

31.03.1998, Webmaster