Even Distribution of Missing Zip Code Level Data
Due to differences in the zip codes that the American Community Survey (ACS) considers to be subregions of counties compared to IMPLAN data, there can be misalignment between raw data and the IMPLAN data that is used to create the demographic expansion figures. This is because the process for Estimating Zip Code Data is separate from the production of Enhanced Demographic Data.
A common issue that this causes involves county-level non-zero values with child zip code level zero values. For instance, a county may indicate that there are non-bilingual speakers of “Asian and Pacific Island Languages”, while the zip codes that belong to that county indicate that there are no individuals that fall into that category. Another example would be a county that indicates there are females in the age ranges of 20 to 24, while no females aged 20 to 24 exist in any of the zip codes that fall under said county according to IMPLAN data.
In the first release of the enhanced demographic data, IMPLAN economists hand-assigned values to cover these differences, most often to the zip code that had the highest population out of the group of zip codes that belonged to the county in question. As of 2021, county-level values are now distributed evenly amongst component zip codes in the event that this mismatch occurs, for first estimates only. This will ensure maximum comparability across years. Note that this is done prior to geographically controlling and RASing all zip code level data.
Improved Household Occupancy Data Estimation
Prior to 2021, Housing Occupancy estimates for zip codes were able to include estimates where households existed in a region, with none of them occupied and some falling under various vacancy categories. Occupancy estimations now incorporate measures to ensure that if households exist in a zip code, there are also some occupied households. It is estimated that about 0.2% of zip codes were affected by this change in both 2020 and 2021 releases.