It can go both ways. Sometimes a nakago-jiri might be slightly truncated and the hamachi/munemachi left in the same position. In other cases, the nakago-jiri might be untouched leaving a longer (sometimes exceedingly long) nakago. I believe there was an early Rai tachi in the Compton collection that had an ubu nakago-jiri and a raised machi that left the nakago with an exaggerated length. All things considered, I believe that machi-okuri is less impactful to shinsa outcomes than modifications to the nakago-jiri, but these things becomes less-and-less important the older the sword is (ie. o-suriage is acceptable for a Nanbokucho blade that is otherwise a juyo candidate, and even moreso for Kamakura-jidai, etc).