Guys,
I was not trying to start a war here. But as far as I can tell, the OP wants SPL capability.
I personally listen at -7 to -10dBref, due to a combination of small children + distortion limitations. I use 93dB mains, and they begin to sound strained above -7dBRef, so counter to the above, sound quality actually is something I seriously pay attention to.
Lemme put it to you this way:
1. IF, again, I say IF Reference Level playback is a high priority for you, low sensitivity speakers are not the best way to go.
2. If you are interested in sound quality over everything else, and do not require Reference Level playback, low sensitivity speakers in a well treated room are a very good option.
3. No matter what speaker type you choose, room treatments are essential to good sound, but that is also a matter of preference vs reference. JBL's Everests will sound like crap in an empty gymnasium to my ears, but may be magic to another's.
If you already have the Statement stuff, build them. I cannot think of a single person that did not think they were amazing, they are world-class.
As for reaching 120dB peaks with 88dB speakers, your 88dB speakers are not providing the 120dB by themselves, they cannot, unless room modes are doing some boosting (SPL meters do not care what frequency the highest amplitude is at). Greater than 115dB peaks are only possible when subwoofer + several speakers are all playing back coherent sound at the limits of the amplitude built into the spec, IF you have flat frequency response at the location your SPL meter is measuring. My old house had a great +15dB peak at 60Hz. Needless to say, Reference Level,was easy to get at that frequency, and 'midbass slam' was quite prominent in any recording. Once EQ'ed, everything was much quieter, and it was far easier to run into distortion limits than SPL limits.
People playing back at Reference with only 100-150W AVR power and 88dB speakers are deluding themselves. By definition there is negative headroom in such a scenario at >3m listening distances for the loudest passages. I would be very curious to test such a system and see how much distortion is present.
JSS