Maybe because he has free medical for hearing aids. 
And, if we're talking about vintage speakers, where some of the oldie acoustic-suspension variety sacrificed efficiency for low-end performance @ higher power inputs, if they recommend, say, 20w input, what happens when you drive them less than that?
This is all about what I suspected, except the talk of transients and clipping kind of goes over my head, especially if I think of a SET volume level on the volume pot. Do these transients that require a lot of power REQUIRE a momentary 100%+ power reserve to be able to reproduce them cleanly?
And does higher WPC usually result in "cleaner" power at lower levels? Maybe I should read a textbook on this.
What happens is that the amp can start "clipping." To be very simplistic: say your amp is multiplying the signal by 10 times. You have a portion of the music that is loud or particularly demanding. Loud = one guitar or drum standing out. Complex = a lot of different instruments/voices at once across the frequencies. When this happens your amp might not be able to multiply the signal by 10x. It only has enough wattage to do say 5x. Well the wave form coming out of the amp gets chopped. At some level the voltage hits the ceiling and it turns into DC. Think of a sine wave. Now flatten the top and the bottom peaks. Those flat lines are DC. And DC is bad for speakers voice coils. Tweeters have tiny ones and are easily damaged this way.
So, the paradox, that speakers are frequently damaged more by amps that have too little wattage.
It's not just about how loud you want to play. If your amp can play the music you like clearly at volumes you want, then all is well.
But sometimes clipping can happen even at reasonable sound levels.
So back to my moderately powered 60 WPC system. From a loudness perspective, I really can't turn the volume all the way up and stay in the room without hurting my ears. But on really demanding, complex pieces of music, it can start clipping at half volume. And since there are exponential curves involved, it would take a MUCH larger amp for me to be able to play those pieces without clipping.
Just trying to come to an understanding about the NEED for high power, vs the DESIRE for high power, vs a sustained "safe", enjoyable listening level.
Lots of assholes and opinions here (equal amount, who'd of thunk?)
My brother used to tell me that a good tube amp sounds better than SS because the coupling of the output transformer as an inductive device is better able to match to the speaker load.
If you have music with high dynamic range, say 20dB peak above average level, that 20dB peak requires 100x the power of the average level to cleanly reproduce it. If you have a 20wpc amp playing at a average output of 1/2 watt and a 20dB peak comes along, guess what? You'd need at least a 50wpc amp to cover that peak. Now, that said, 20dB peaks are not all that common in contemporary music.
So, yeah, higher WPC may or may not result in cleaner playback at lower levels, it depends on several factors at play. There isn't really a strict yes or no answer that fits every case.
I just listen at married-guy-with-kids SPL.

.
Peaks in a good recording will be 12-15dB louder than its average loudness (or even more). Each 3dB requires a doubling of power. This is true no matter the volume (loudness) you are listening at. If your amplifier cannot deliver the necessary power, it will clip...which simply means that it does not have a enough power to fully reproduce the signal. In turn, this distorts it. Minor clipping is not obvious, but it quickly becomes more apparent.