Matching IPUMS-CPS monthly files

Good morning,

Probably I did not explain myself properly. I am trying to match montly IPUMS-CPS data from 1989 to 2010 (jenaury with february, with march, with april, etc.). This can be easily done from 2004 to 2010 by

  1. collecting hrhhid and HRHHID2 from the NBER March supplement (with a sequential merge with IPUMS March file) and merging them to the IPUMS March files,

  2. and then by applying the (modified) algorithm by Madrian, B. and L. J. Lefgren (1999)

because indeed NBER MArch Supplements contain exactly the variable hrhhid and HRHHID2.

HOWEVER, there is a problem when trying to merge the other IPUMS March Files (not comprehensive of hrhhid and HRHHID) with the NBER March supplements for the years 1989-2004. In fact for these years the NBER files have others identifiers(e.g. h_idnum, etc.), but I need always hrhhid and HRHHID2 because these are the identifiers used in IPUMS non March files. Therefore I need hrhhid and HRHHID2 to implement the match of all IPUMS monthly files

NOW MY QUESTION IS: how can I build hrhhid and HRHHID2 from the identifiers used in the NBER MArch supplement for years before 2004?

this would allow me to match all IPUMS cps monthly files available on your website

Any suggestion is apreciated

Thankyou

Ric

Sorry for the misunderstanding on your previous question. Unfortunately, there is not a simple solution to matching across months going back to 1989. The Census typically creates the HRHHID variable by combining a geographical identifier with a completely random number. As a result, it is not possible to simply reconstruct these variables for earlier samples in a manner that would facilitate matching with pre-existing HRHHID variables. Similarly, for HRHHID2, the NBER components (HRSAMPLE, HRSERSUF, and HUHHNUM) are not available prior to January 1994.

However, I can point you to a couple resources that may be of use. First, this paper discusses a method for creating unique identifiers for CPS persons and households back to 1979. These identifiers will eventually be available as a part of IPUMS-CPS. Second, Section III of the Census’ “How to Link CPS Public Use Data Files” document provides the identifiers necessary for matching March supplements with non-March supplements back to 1994. Additionally, the previously provided NBER documentation provides a link to the identifiers necessary for matching basic CPS and March CPS files across months.

I hope these resources are of some assistance.