I use CPS ASEC 1979-2015 for my analysis. Since I need baseline characteristics of individuals, I link individuals from CPS ASEC in year X to CPS ASEC in year X+1. In my analysis I use individuals observed in year X+1 and their characteristics in year X. Which weights can I use for my analysis? Can I just use the normal ASEC weights or do I have to make some adjustments?
There is limited guidance on weighting longitudinal analyses of the CPS. You may be interested in LNKFW1YWT, a forward-looking longitudinal weight for linking the CPS for two adjacent years. This might be appropriate if you are interested in individuals observed in year x+1 and leveraging characteristics from year x. ASECWT is not a linking weight and therefore it does not adjust for non-linkage, but it is possible that it is appropriate depending on your application. For more information on longitudinal weights in IPUMS CPS, including Stata code used to generate the weights that can be used to generate backward-looking weights, see this page on Linking and the CPS.
LNKFW1YWT is not available for years prior to 1989. Will the weights be available for these years?
The stata replication files are only available for longitudinal weights for two adjacent months. Is it somewhere available for two adjacent years?
LNKFW1YWT is created based on CPSIDP. Back when it was originally constructed, CPSIDP was only available going back to 1989; it has since been extended back to 1976 (see Flood et al., 2020 for more information). The IPUMS CPS team is planning on extending the availability of LNKFW1YWT to earlier years, which should become available in the next couple of months.
Stata replication files available in the CPS Linking Documentation are only available for LNKFW1MWT, but this code is intended as a baseline that users can modify to suit their needs.