Per capita income at Place/Remainder-County-County Subdivision level


I am working on an analysis of per capita income at the Census Place level. I am trying to get ACS 2019 5-year estimates data with population and per capita income for each Census Place and also for the remainder of each county. Thus the total population would sum to the total for the country overall (I believe this is 328,016,242 in the 2019 5-year estimates).

It appears that the geographic level I need is Place/Remainder by State–County–County Subdivision (NHGIS ID: PLACE_070; NHGIS Code: 070). When I look up the per capita income variable (Table B19301) in the Data Finder, it says that it is available at this geographic level. However, when I download this data, the column for per capita income is completely empty. The column for total population (Table B01003) does populate correctly.

Is there something I am doing wrong? If I instead download at the Place by State geographic level (NHGIS ID: place; NHGIS Code: 160), or at the County by State–Place level (NHGIS ID: county_155; NHGIS Code: 155), the column populates, however these levels appear to leave out any areas that are not within a Place, since in both cases the population sums to just 244,451,098.

I believe I can create the data I want manually by combining the County by State–Place data with the standard County data to get information on the non-Place areas within each county. But before I start on that I wanted to check whether it is possible to download the Place/Remainder by county data directly, since the documentation suggests it is available.

Separately, is there a list somewhere that describes the geographic levels in detail, perhaps especially which ones include the entire country and which ones do not?

Thank you!!

Hi Robert,

I just faced this situation for a project I completed in April/May. I too was hoping to use the Place_070 geog. level to get at per capita income for places and the county remainders, but it appears that the per capita income isn’t published for that particular geog level. My colleague, Jonathan Schroeder, may also chime in, but the ACS doesn’t always include data for for particular table x geog. level combos, especially those where some of the geog units can be quite small (e.g., the county remainders may include too few people to publish a reliable estimate).

I ended up combining the county_155 geographic level with the county geographic level (NHGIS ID: county; NHGIS code: 050) level to compute the per capita income. I acquired the sex by age table (B01001) and the aggregate income table (B19313. Aggregate Income in the Past 12 Months (in 2019 Inflation-Adjusted Dollars)) for both of these geographic levels.

I then calculated the # of people 15 years and older from B01001 for both geographic levels.

Next, I collapsed county_155 by county, summing the population age 15 and older and the aggregate income. This yields one record per county and the columns represent only those living in places within the county.

Then, I merged the result of the collapse with the county data. I subtracted the collapsed values for population age 15 and older from the county total, and I subtracted the collapse values for aggregate income from the county total. This difference represents the # of persons age 15 plus and the aggregate income for those living in the county but NOT in a place in the county.

Then you can divide the aggregate income values by the # of persons age 15 plus to get the per capita income for the County Remainder. You can use the data for places directly from the county_155 geographic level.

This is definitely a bit of a workaround, but as far as I can tell, it is a way to get measures for the county remainders.

  • Dave Van Riper

Dave is correct that ACS suppresses data for some tables for some geographic levels. The NHGIS filtering system doesn’t capture all of the nuances within datasets, so there are (unfortunately) cases where our system will show that a table is available for a geographic level even when the table contains no data for that level.

I don’t know of a public listing anywhere that identifies exactly which tables are available for which levels or vice versa. We construct a complex, internal version of this listing when processing ACS data, and in the more extreme cases, where a table provides data for only a small number of levels, we add bracketed text to the NHGIS table label to indicate the table’s valid levels (e.g., our label for table B29001: “Citizen, Voting-Age Population by Age [Data for nation, state, county, tract, block group, place, congressional district, and state legislative district levels only]”). But we don’t add this info in cases where the table is available for most of a dataset’s levels as it would make table label very long.

I think the strategy Dave outlined should allow you to compute per capita income for place remainders without too much trouble. I’d just note that, to my knowledge, standard measures of per capita income are not limited to working age population (age 15 and over) in the way Dave describes. E.g., according to ACS documentation, “Per capita income is the mean income computed for every man,
woman, and child in a particular group including those living in group quarters.” So you may want to adjust his instructions and obtain total population (available in table B01003) and use that as your denominator to produce a standard measure of per capita income.

Thank you both for the super helpful and quick responses! That makes sense on the challenges with the documentation, and this approach will work for getting the information I need.

I really appreciate this!!