IPUMS CPS is a rotating panel. If I use two consecutive years (2012-2013) March ASEC data, using the same sample of respondents by getting rid of MISH 5678 respondents in 2012 and MISH 1234 respondents in 2013 (should be close to the same sample after), can I identify the respondent in 2012 and 2013 (by some serial number of characteristic) in order to run a panel regression by tracking individuals from 2012 to 2013? The serial numbers are different every year I believe? Is there any other identifier over time?
SERIAL is generated by IPUMS for each sample and cannot be used to track households across samples. Instead, you will want to use HRHHID and HRHHID2 to link ASEC households between consecutive years and include LINENO, SEX, and AGE (allowing age to increase by one) to link individual respondents.
Hope this helps.