Historic-to-modern MSA crosswalk

Hi all,

I’m looking to study the evolution of some specific outcome variables in metro areas throughout the 20th-21st centuries. Of course, MSA definitions change over time. I figured the way I would proceed is by fixing my geographic MSA units over time by using a single, contemporary MSA definition.

I wanted to know if there are any pre-constructed crosswalks from earlier geographic units (e.g. earlier MSA’s, counties, cities, etc.) to modern (eg. 2010) MSA definitions. The “comparability” of the metarea variable does have a county mapping across decades to contemporary MSA, but perhaps there is something that is a bit more straightforward or that is in more common use that I’m unaware of.

Just to be very clear, for example: I’ll download the 1930 (full) census, which includes geographic household variables for county, city, state, and metroarea. Is there a dataset that would very simply map any of these geographic units to a 2010 metro area (MSA)?

Sorry if this already has an answer—I haven’t been able to find anything more direct myself.

There are no pre-constructed crosswalks from earlier geographic units to modern MSA definitions. You can create one using the following steps:

  1. Acquire the Census Bureaus crosswalk for counties to MSA. The 2013-present delineation files cover county to Core-Based Statistical Area (CBSA) relationships, and historical delineation files cover county to MSA/CBSA relationships for years prior to 2010.
  2. In your 1930 complete count extract, make sure you have the variable COUNTYNHG (County NHGIS Code) included in your request. You will use the COUNTYNHG codes to merge onto the delineation files.
  3. Generate a list of unique COUNTYNHG values in your extract. * Create a COUNTYNHG value in the delineation file by concatenating the following values together: “G” + state fips code + “0” + county fips code + “0”
  4. Link the two data files by COUNTYNHG. Mismatches (e.g., COUNTYNHG records from the microdata without a match in the delineation file; COUNTYNHG records from the delineation file without a match in the microdata) will require adjudication. You will need to do some research to reconcile the mismatches. One way you can start is to acquire 1930 and 2010 county shapefiles from IPUMS NHGIS and do a spatial overlay to establish a spatial relationship between the two decades’ counties. This may help you fix up some of the mismatches. Note 1: in the shapefiles, the GISJOIN field contains the same codes as the COUNTYNHG field in the microdata. Note 2: I strongly recommend getting the 1930 county shapefiles with a “2008 TIGER/Line +” basis. We realigned the geometry in our original 1930 county shapefiles to match the 2008 TIGER geometry, which is similar to the 2010 geometry.

This linking strategy is mostly based on county codes and does not take county boundary changes into account. I strongly recommend using those 1930 and 2010 county boundary shapefiles to investigate whether county boundary changes have occured in MSAs.

1 Like

Thank you so much for sharing you experience on this topic! This is certainly really helpful going forward—I also definitely agree with the recommendation to look at historic county boundary changes.

Thank you again!