I think it's the combination of both. The lower mass and glued in reflex seem like they would account for the difference.
I've never liked Baker's claim that Perry reflex relieves the belly of strain. Bows cut apart after use show that belly was under just as much strain and it's just hidden by the back pulling it back when unstrung.
Let me pick a worm out of that can 
Depends on how you look at it.
if your bow is 10% lighter for the same draw weight, you have less wood to strain. So surely it must resist straining better, if it requires less wood for the same amount of force that's applied. In the end, it's strained the same amount (takes as much set), but since there's less of it, you have less mass, and hence there's an advantage that is, indirectly, the result of perry-reflexed bellies handling strain better.
I look at it as more of the mass is actually being strained rather than carrying dead weight. It just doesn't seem like the belly is being relieved judging by actual set.
It kind of seemed like the early theory would have the belly be straight if it were removed after use.
