I’m not entirely certain I’m understanding your research project correctly, so forgive me if this answer is a bit unhelpful.
A tricky detail is that the ACS and US Census surveys are all sampled at the household level. Therefore the FAMUNIT, RELATE, and FTOTINC variables only provide family interrelationship information for individuals within households. Once a child moves and creates a new household, it is impossible (with the public use data) to identify an inter-family relationship across households. You could perform your analysis within households and analyze inter-generational wage changes of families within households, but you’ll be excluding all the families who live in two different households.
One alternative is to use the IPUMS Linked Representative samples, from 1850-1930. This resource provides data samples for 7 pairs of years: 1850-1880, 1860-1880, 1870-1880, 1880-1900, 1880-1910, 1880-1920, and 1880-1930; each containing three independent linked samples: one of men, one of women, and one of married couples. Each linked sample contains individuals who can be identified in two distinct census years. I’ve seen researchers use this data to analyze inter-generational dynamics in a very useful manner. More information on the IPUMS Linked Representative samples is available here.
I hope this helps.