I am able to replicate the official poverty variable. But I want to replicate IPUMS CPS POVERTY variable. I am dividng FTOTVAL by CUTOFF and then recoding to see if I can match the 4 categories of POVERTY. It is not matching and in fact way off. What else should I do? Thank you!
There are a lot of reasons that could potentially explain the discrepancies you are seeing. I suggest using OFFTOTVAL and OFFCUTOFF instead. Remember to limit by OFFPOVUNIV.
This page should be useful: https://cps.ipums.org/cps/poverty_not…
I would go with OFFTOTVAL and OFFCUTOFF, because as I mentioned I am able to replicate the official poverty measure. However I have informtunately done a bunch of work using the POVERTY variable and now we need more additional threshold cutoffs i.e. below 200% poverty. If I new switch gears and use the official, the total #'s will not correspond to what I have already submitted using the POVERTY variable categories.
We’ll need a little more background to help you. What years of data are you looking at? What rates are you finding with FTOTVAL by CUTOFF?
I’ll also encourage you again to read this page in depth: https://cps.ipums.org/cps/poverty_not…. The impetus for creating OFFPOV was the fact that POVERTY does not match official poverty rates. I suspect that many of the issue with using FTOTVAL and CUTOFF will be related to what we found with the OFFPOV work.