I doubt the calculated efficiency will be much different, other than the no-load
losses (control loop supply current).
However there are some losses which loop stability (or more properly loop
instability) can really magnify. If the loop is at all oscillatory then power is
sloshed from supply to load, then load to ground, shuttling a lot of energy
to no good end. That will kill efficiency really quickly. Since Zout is low even
tens of mV oscillation at the feedback node can be a lot of current wasted.
Your controllers' phasing of HSS and LSS is also critical and might vary between
the two schemes; soft switching or the amount of undershoot in the LSS are
important to efficiency as well.
Whether these appear in simulation and how that relates to reality, is about
the quality of modeling, parasitics, etc.