This experiment combined 31 types of administrative record and third-party sources to produce 2020 population estimates with the same reference date, April 1, 2020, and within the same timeframe as the 2020 Census of Population and Housing. The sources and methodology are designed to improve coverage of historically undercounted population groups.
The administrative census national population estimates are 1.1 and 2.5 percent higher than the 2020 Demographic Analysis high and low estimates, respectively, 2.3 percent higher than the 2020 Census count, and 2.9 percent higher than the vintage-2020 Population Estimates Program estimate. Our analysis suggests that the higher administrative record census estimate can be explained by inclusion of more non-U.S. citizens, especially those with unknown legal status.
The report identifies several improvements critical to developing the Census Bureau’s administrative record infrastructure to facilitate higher-quality administrative record-based population estimates, especially regarding locational accuracy.