It depends. I own both balanced and single-ended editions of otherwise identical interconnect, in two different cable types, all of which are 1m length. Between my internally-balanced DAC and my internally-balanced preamp, there's no question in my mind that going balanced makes a small but musically worthwhile improvement, at least in my system. I tested this by doing both hook-ups simultaneously and switching between them instantaneously via the remote from the listening chair with the relative volume levels preprogrammed to match. In audiophile shorthand, the balanced connection lends the music a distinct bit more 'palpable presence' - in other words, it sounds a little closer to life than it does via the single-ended connection, presumably because it's slightly less degraded along the way, and because the connected balanced circuits can function differentially without the interposing of lossy adaptors or phase-splitting stages.
On the other hand, between the aforementioned internally-balanced preamp and my non-balanced power amps, there seems to be a slight loss switching to a balanced interconnect. This may not be too surprising, since it was predicted by the amps' manufacturer, but of course I felt the need to find out for myself anyway. I would not characterize the difference as being quite as meaningful (in the other direction) as was the improvement from source to preamp, which makes sense as it probably only has to do with the balanced input signal's being routed through a bit of extra contacts/solder joints/wire on its way to ending up at the same place and in the same form as a single-ended input is from the start, with no other theoretical advantages accruing to the single-ended connection.
These results lead me to two tentative, anecdotally-supported conclusions: 1) Short wire runs are no reason to automatically dismiss balanced connection in lieu of direct experimentation; 2) Whether or not the sending and receiving circuits involved are truly balanced, rather than simply providing XLR connection to a single-ended circuit for convenience's sake, will have a lot to do with the existence of any potential upside. And then there's the issue of the wire itself - from what I understand, a lot of interconnect models (including the ones I used) are internally no different between the balanced and unbalanced versions except for their termination, but not all, and it stands to reason that if you're talking about one of those models that's actually differently constructed for each version, then all bets would be off.
On the other hand, between the aforementioned internally-balanced preamp and my non-balanced power amps, there seems to be a slight loss switching to a balanced interconnect. This may not be too surprising, since it was predicted by the amps' manufacturer, but of course I felt the need to find out for myself anyway. I would not characterize the difference as being quite as meaningful (in the other direction) as was the improvement from source to preamp, which makes sense as it probably only has to do with the balanced input signal's being routed through a bit of extra contacts/solder joints/wire on its way to ending up at the same place and in the same form as a single-ended input is from the start, with no other theoretical advantages accruing to the single-ended connection.
These results lead me to two tentative, anecdotally-supported conclusions: 1) Short wire runs are no reason to automatically dismiss balanced connection in lieu of direct experimentation; 2) Whether or not the sending and receiving circuits involved are truly balanced, rather than simply providing XLR connection to a single-ended circuit for convenience's sake, will have a lot to do with the existence of any potential upside. And then there's the issue of the wire itself - from what I understand, a lot of interconnect models (including the ones I used) are internally no different between the balanced and unbalanced versions except for their termination, but not all, and it stands to reason that if you're talking about one of those models that's actually differently constructed for each version, then all bets would be off.