Note: [1] I am grateful for the comments and suggestions made by Douglas Bracy, Paul
Coomes, Joseph Duncan, David Kendrick, Michael McCracken, Robert McGuckin,
and Jack Triplett. I have also benefited greatly from conversations with
Patricia Renfro on Search Engines and their application. However, all errors
that remain and interpretations made are my own.
Abstract: This paper considers the current and prospective development of economic database
systems, thematically complementing a 1980 paper by the author. First created
for use with macroeconomic data, often in order to support the construction and
use of econometric and other models, these systems have been applied in this way
and more broadly since the 1970s. However, during the past decade, microeconomic
systems have also begun to be developed, particularly for use with longitudinal
data sets. With the advent of the Internet, using next generation systems, the
potential now exists for the widespread, cost and use-effective distribution of
various types of economic and social data for research and other purposes. This
paper examines some of the requirements for the appropriate development of these
systems. It stresses the absolute need for improved data documentation and
advocates the machine distribution of documented data as a standard procedure.
It evaluates proposals by Kendrick and David that economic database systems
should be organized according to the Relational Database Model, a computer
science concept originally formulated by E.F. Codd. The appendix presents and
describes the widely used TSD economic data transfer format, which permits
documented inter-system time-series data transfers.
Keywords: Computer, database, DBMS, economic, longitudinal, relational, statistics, temporal, time series
