Hi,
I have been asked to compare three iterations of Table B19001 (Household Income in the last 12 months), using the 2009-2013, 2014-2018, and 2019-2023 American Community Survey (ACS) 5-year data files.
The project requires reporting the weighted number of observations for each income category, the margin of error associated with each category, and a percentage estimate along with its margin of error. Due to these requirements, I am unable to use the ACS comparison table (Table CP03). Additionally, Table DP03 compares two points in time, and my project requires comparing three points in time.
How would one adjust the number of observations for each income category to reflect values in 2023 dollars? Given that each iteration of Table B19001 uses the same categorical boundaries ("Less than $10,000", "$10,000 to $14,999", etc.), is adjusting the counts to create 2023 dollars even appropriate?
I know how to adjust dollar values (such as median income) to a constant value using the Consumer Price Index (CPI), but adjusting categorical boundaries seems to be a different operation.
Has anyone encountered this problem before? How should one proceed?
TIA
AB
Hi Adam, is using the ACS microdata an option? If so, I think you could inflation adjust the income values at the household level and then assign them to a comparable set of income categories for each time period. The results wouldn't match those from Table B19001 exactly but should be close, at least for larger geographies.
Thanks @Mark
Using the microdata isn't an option. I need this data at the county and place levels. However, your comment confirms my suspicions that doing the adjustments I was considering is not possible. I'll have to use the published tables, noting in the background documentation that the underlying values are not adjusted for inflation. I suppose this is why we have documentation in the first place.
Cheers -- AB