I’m trying to extend the CPS IPUMS dataset with the NBER outgoing rotation files (http://www.nber.org/morg/annual/), because the former only has data on hourly wages until 1989, the latter until 1979. I’m using the basic monthly IPUMS CPS dataset.
I found this post: blogpostsaying that I should use hrhhid (hhid in NBER), hrhhid2, huhhnum (hhnum in NBER), hrsample, hrsersuf and lineno as keys in the merge, but for these early ORG samples, hrhhid2, hrsample and hrsersuf don’t exist and the other three variables don’t uniquely identify observations in either the CPS or the NBER files.
This documentsuggests that I should use hhid, hhnum, lineno and combine these with minsamp (mish in CPS) and intmonth (month in CPS) to use them as keys in the merge. But even then, variables are not uniquely identified.
I’ve taken a bit of time to look into this myself and also find that these variables (even with month in sample, month, and year included) do not uniquely identify observations in the NBER MORG samples from 1979 through 1989. This being the case, I am not able to provide much guidance on linking IPUMS CPS outgoing rotation group data to the NBER MORG samples during this time frame.
This detail is related to a larger ongoing project with IPUMS CPS that aims to allow full comparability and linking of the outgoing rotation groups all the way back to 1979. This project is not yet completed, however, and so there is not much additional detail that I can offer. Regarding unique identification and the perhaps incorrect documentation, I’d suggest reaching out to Jean Roth at jroth@nber.org.
Hi Jeff, did this project ever get completed? I find that year “hrhhid lineno misg month age sex statefip” do a reasonable job of unique id’ing both ipums cps basic and nber org (e.g non unique ids using these vars are much less than 1% of monthly samples) except in 2019 where it fails for ipums cps basic. Any idea why?
I believe the project that Jeff was referring to led to the extension of CPSIDP back to 1976, and is detailed in the paper by Flood et al. (2020). I’m not sure about the MORG linking in 2019, but you should be able to get all the info available in the NBER MORG files through IPUMS CPS.