In visions of future government statistical systems, administrative records will be used widely. Demand for information will increase beyond the ability of surveys and censuses to meet it. Administrative records will become easier to get and process, enabling more and cheaper estimates with less burden on respondents. But administrative records are a by product of systems that we currently neither fully understand nor control, so many concerns arise about our ability to use them and the quality of estimates based on them. Last January, the authors conducted a review (Marquis, Palacios, and Wetrogan 1996) and concluded that many of the Census Bureau's programs would benefit by creating a national database of people linked to families, addresses, and geographic areas. This paper raises some orienting issues concerning a national population database, provides some early results related to a few of those issues, and discusses next steps for moving ahead with database building.
