Technically, CPSIDP links the respondent appearing on a designated household roster line number (LINENO) across samples by assigning a unique CPSIDP value to this line number. Since the original person may leave the household, the respondent appearing on a given line number can be replaced; however, the CPSIDP value will remain constant for a maximum of 8 months (i.e. the length of time a household can remain in the CPS survey). As a result, it is important to verify your CPSIDP linkages with the AGE, SEX, and RACE variables, allowing for AGE to increase. Due to the complexity of linking persons and households across CPS samples (e.g. migration, mortality, non-response, and recording errors), erroneous links can still sometimes occur among those that match based on CPSIDP and demographic characteristics. Since the IPUMS-CPS linking variables are built using the original CPS variables, any errors in the source data will be reflected in CPSIDP. Ultimately, it is up to the individual researcher to determine the acceptability of the linkages made using CPSIDP.
Hope this helps.