I would suggest running pink noise (easy for a voltmeter to measure) at a good loud level. Using an AC voltmeter - a simple analog scale multimeter will be fine; a nice true RMS voltmeter even better - measure the voltage across the speaker terminals. Power, simplified to ignore power factor and reactive loading, is E *E /R, that is, voltage squared divided by resistance. The resistance may be considered the average impedance of the speaker. So, if you read 5 VAC across a 4 ohm speaker's terminals, power is 5*5/4, or 6.25 watts. I think you will find that 5 volts is pretty darned loud. By the way, a 220-watt amp will make 30 Volts across that 4 ohm load.
The trouble is, this does not account for musical peaks. I do not know how the peak-and-valley nature of musical dynamics translates into steady-state RMS equivalent. If we consider the dynamic range of music to go up to twice as loud on the peaks as your average listening level, you will be looking at a 10 dB increase, which also happens to be 10 x the power - so your 6.25 watts average needs to make 62.5 watts for the loud bit. But how many loud bits occur, and how long do they last? I'm not really sure that a millisecond-length musical peak matters when compared to the continuous RMS power capability of the amp - up until it clips against the the rail voltage, anyway.
Another thing you might try is looking at a fast meter - VU response - or a peak-reading meter during a loudly-played piece of actual music. See what the voltage peaks up to during the loudest bit, if you can look fast enough.
Some amps will easily make a 30-volt output (225 watts @ 4 ohms), but will overheat if driving a 4-ohm load at 30 volts for an extended period - multiples of minutes, that is. After all, 30 volts across a 4 ohm load will run 7.5 amps of current through the load. That's a LOT of current running through an amplifier's output stage. Your speakers would melt before the amp in that case, though - long before the amp. I think that's why amplifier makers no longer use speaker fuses (an anachronism from the early 70's, when nobody had made high power amps before). A speaker, tweeters especially, will conveniently blow to protect a fuse, and not the other way 'round.
This is pretty simplified and un-technical, so I may have ignored some of the complicated engineering bits, but I hope this helps you measure what you need.
I'm kind of interested in this myself, and would really like to read others' ideas on the subject.