I can't envision any reason as to why the match or mismatch of IC and SC bandwidths (essentially the upper limit of the range of frequencies the cables can pass without significant attenuation/rolloff) would be of any significance.
Also, the upper limit to the range of frequencies an IC can pass will vary significantly depending on the output impedance of the component which drives the cable. And in most cases the bandwidth of both types of cable will be (and should be) much greater than the bandwidths of our ears, the speakers, and usually some or all of the electronic components. Which means that what those cable bandwidths are won't matter, because other things will limit the overall bandwidth.
That said, though, I can't speak too knowledgeably about MIT cables in particular. In part because I've never seen a clear and technically meaningful definition of what Mr. Brisson means by "articulation poles," and in part because I've never seen a schematic or other technical description of what is in his network boxes.
Regarding your post just above, I agree. I would add that even in the case of many variables that can be identified, such as many of as those I listed in my post dated 12-15-12
in the thread I linked to earlier, the resulting sonic consequences, if any, will in many cases have little or no predictability. The effects of electrical noise and RFI on the sonics of a component being one example.