"This could occur if the individual base layer fees are too high"
Tail emission would suppress fees growth, that's clear. The only reasonable (because free market version) of tail emission is to add another 210k blocks delay to the next halving - only in case if the network difficulty hasn't recover during four years long period after the previous halving.
P.S. I'm not in a crazy rush for this change to happen, but in case of first "destructive halving" - I'm gonna be... :)