Migrate5 reports migration across unharmonized administrative units. That might explain the unexpected results that you’re getting since there may be cases in which units that existed 5 years ago do not exist in the current census or where new units were created. To mitigate this issue, you will need to use a map of the unharmonized geographic units to determine what changes were made to that unit (i.e. which administrative unit it was combined with/separated from).
There are a couple ways IPUMS can help you in this process. The first is that IPUMS does provide harmonized migration data on first-level administrative units in GEOMIG1_5.
Second is that though the process of harmonizing data on second-level units is still on-going, there are a few countries for which this process is complete. If you reach out to us at ipums@umn.edu and the country is available, I can share translation tables that would help you in deriving second-level administrative units for migration variables that are consistent over time.
This is a very complicated question, so I hope this answer helps!