i've posted this reply a few times, so if you want larger details, you might want to search these forums.
it comes down to basic signal/systems. basically each filter has an effect on the magnitude and phase of the input signal.
thus if you have a 12dB/oct slope and a -18db/oct slope, you get a -30dB/oct slope.
but what about the fine print?
ok. when people make filters, they usually try to aim for something useful. lets take the example of 1st order lowpass filters that are cascaded. the first filter will give a -6dB/oct response at, say 50hz. the second will give a -6dB/oct response at 500hz.
for the above example, the filter has a -12dB/oct final slope, but still isn't too useful. the slope is only -6dB/oct in the 50-500hz band, which is probably the "important" band. the 500hz LPF really doesn't do much.
the second fine print is that people define the cutoff as the -3dB point. if filter A is -3dB @ 80hz, and filter B is -3dB @ 80hz, then the combined filter is -6dB @ 80hz. hmm. seems the cutoff point, as we've defined it, has moved.
when you comine filters you will get the high final slope, but the slope "of interest" probably won't be as steep. further, the "cutoff point" moves, and you probably won't know what the phase response of the combined filter will be. for these reasons, its better to use just one filter if possible.
edit -- as for sound, steep slopes don't equate to magically good sound. choose what works for the speakers ect...