We frequently use the missouri data center MABLE/Geocorr to get PUMA to Metro crosswalks, but I see that their 1990 crosswalk tool does not correspond to anything in a later geography. This leaves me wondering how I can get from 1990 PUMAs to some metro definition that could be consistent with 2016 metros (as crosswalked from 2010 PUMAs using MABLE/Geocorr). Would be grateful for any insights / advice!
First, you can get a crosswalk from 1990 PUMAs to other 1990 geographic units through the IPUMS USA 1990 PUMA Definitions page. (Note that the table on that page lists crosswalks by state, but there are also nationwide crosswalks available through the links just above the table.)
Second, you can get delineation files for current (2017, 2015 or 2013) definitions of metro areas (i.e., core-based statistical areas or “CBSAs”) through this Census Bureau page.
In _most_ cases, the 1990 counties in the 1990 PUMA equivalency files correspond directly to the current counties in the CBSA delineation files, so it should generally be possible to bridge directly from the 1990 PUMA crosswalk through the CBSA delineation files to create a 1990-PUMA-to-current-metro-area crosswalk.
Another Census page lists Substantial Changes to Counties since 1970. A couple changes that I know would have an impact on the 1990-PUMA-to-current-CBSA crosswalk are: the creation of Broomfield County, Colorado in 2001 (which splits off parts of the Boulder and Greeley metro areas into the Denver area), and the change of the name of Dade County, Florida, to Miami-Dade County in 1997, which also involves a county FIPS code change from 025 to 086. There may be other relevant changes, and unfortunately, I don’t have a simple answer for how to deal with them–especially cases that involve county boundary changes, as with Broomfield. There are some fairly simple approaches, but the simplest options are not necessarily accurate.
Note also that the Census Bureau’s 2016 ACS summary files (as provided by IPUMS NHGIS) use 2015 CBSA definitions, but if you’re obtaining any microdata from IPUMS USA, you might prefer to use 2013 CBSA definitions, as used by the MET2013 variable in IPUMS USA. (We do not have an equivalent “MET2015” variable.) As you may know, MET2013 supplies 2013 metro area codes, where possible, back through the 2000 microdata samples. We have plans to extend MET2013 back through 1990 and possibly 1980, but unfortunately for your application, it may be a year or two before we get to that!
Feel free to respond with more questions if you have any.
Dear Jonathan,
in this post from 2018 you mention plans to extend MET2013 back through 1990. Since I would need such a crosswalk as well, I wanted to ask what the current status of this project ist.
Thanks,
Andreas
Hi Andreas,
Unfortunately, the status hasn’t changed. (I spoke too soon in my last post to suggest it’d only be a year or two!) We try to make improvements to IPUMS USA geographic variables around once a year, and we’ll be due for more updates in ~6-9 months, but our to-do list is long, so I can’t say whether this update will be in our next batch. Your inquiry is helpful, though… We do try to prioritize updates that are in demand!
Hi Jonathan (@JonathanSchroeder)
Any update on this please? Would be massively helpful. I have manually constructed a crosswalk based on county based metro definitions, but there are so many metroareas that have increased in geography, and then there are new metros as well. Counties are also not always identified in ipums, so there is that problem as well.
@Umair_Ali, we have unfortunately not gotten to this yet. It’s at the top of the to-do list for IPUMS USA geography, so I’d guess we’ll fit it in this year sometime, but it will be several more months at the earliest. Thanks for repeating the request.
Hello,
Just doing the yearly check-in on this. There’s clearly an immense need for this crosswalk based on this thread. The OMB coding that IPUMS uses for CBSAs seems to come from an alternate set of CBSA coding which does not perfectly conform to https://www.cms.gov/Medicare/Medicare-Fee-for-Service-Payment/AcuteInpatientPPS/Acute-Inpatient-Files-for-Download-Items/CMS022637 for very large areas like New York, Chicago, LA, etc. This makes this crosswalking very difficult to do manually. Having IPUMS provide it would be a massive lift that enables policymakers, policy researchers/analysts, and social scientists to do long-term studies of demographics and housing.
Any info. you can provide would be great. Thanks for the hard work you all are doing over at IPUMS.
@Chris_B, I’m afraid I’m becoming a bit of a broken record in this thread… We still have no near-term plans to extend MET2013 back to 1990. We still hope to do that, but I won’t set myself up again by suggesting a timeline. I agree with all the points you make about the potential value, but unfortunately, we’ve had no staff principally dedicated to this type of work, and it’s continually been difficult to carve out time away from other projects here. Your post has helped again to remind us of the need!
1 Like