You are correct that YNGCH refers to the age of the youngest own child of an individual. What you are observing appears to be the result of the following three things:
(1) In many of these cases, AGE itself has been allocated. You can identify these cases using QAGE.
(2) YNGCH includes step-children and adopted children as well as biological children. In many of the cases where YNGCH is greater than AGE, it appears that there is a step-child involved and these could be legitimate differences. For example, there is a household where the head of the household is female/63 and the spouse is male/34. In this household, there is a child who is 45. Using PEMOMTYP and PEDADTYP, you can see that this child is the biological child of the head of the household (female/63) but step-child of the spouse (male/34), which appears to be a legitimate composition.
(3) YNGCH is an IPUMS-derived variable using IPUMS-derived family interrelationships. As such, information in YNGCH may differ from family information that comes from just the Census family definitions.
I hope this helps.