Geographic adjustments

I’ve been digging into the spm geographic adjustments, and I’ve noted that the base geographic adjustments reported by Fox (2018) for 2017 – in the table " SPM Thresholds by Metro Area" The Supplemental Poverty Measure: 2017, and also in the underlying research data set (variable spmu_geoadj) do not match the base thresholds in the IPUMS CPS data set (spmgeoadj) for specific geographic areas (e.g., Oregon metro, Oregon nonmetro). There are seven spmgeoadj values for the base threshold of each housing tenure type in both data sets, but that don’t match, which doesn’t make sense to me, since they should be based on the same underlying variables (index of median gross rents, share of threshold devoted to shelter and utilities, and BLS FCSU thresholds for each tenure group).

Can you explain why this is the case? Did IPUMS use a different underlying median gross rents index and/or share of threshold for the various housing tenure groups? Do I need to worry about this?


This is an interesting observation. According to our translation table, our integrated variable should be identical to the source variable. It could be the case that you are comparing the 2017 table with the 2017 ASEC data, but since ASEC data report on the previous year, the comparison should really be with the 2018 ASEC data. If the issue seems to persist, please feel free to send additional information directly to and we can help.

Thanks for the quick response. I just compared the 2017 IPUMS ASEC data with the Census “2016” file, and they match perfectly! So I presume that when the IPUMS 2018 ASEC data are released, they’ll match the 2017 Census data and file.

1 Like