I hear differences between PC's on most of my components, some more than others. Do I think these differences would be revealed in a blind test? Could be that they wouldn't: I personally am of the opinion that blind testing is a very good way to *obscure* fine sonic differences, not to reveal them. But I also believe that the placebo effect is quite real, and can be about equally confounding to sighted test results as is the deliberate confusion intoduced by the blind ABX methodology. Still, although I have to admit that some of the differences I've thought I've heard have been very subtle, elusive, or tough to repeat with consistency, others have seemed unmistakable and unchanging.
Anyway, one of the simplest explanations for why the last few feet ought to matter is the question of shielding. A theory goes that a large portion of the EMI radiation the incoming power is exposed to is actually generated by the system components themselves, therefore making shielding an issue between the wall and the gear. Since stock PC's are generally unshielded, if this supposition is true then it could help explain there being an audible difference. This theory might also be turned around: A stock, unshielded cord might be acting as a 'broadcast antenna', radiating 60-cycle noise that could interfere with nearby components.
In fact, a lot of the potential differences could be explained in basically similar terms - that is to say, that the aftermarket replacement cord is doing something competently that the stock cord did poorly. For instance, if the stock cord distorts the incoming power due to spurious internal strand resonances, this might have a detrimental effect just as it could in a speaker cable or interconnect. Or the stock cord might serve as somewhat of a 'bottleneck' during heavy demand (maybe being too light in guage), current-limiting or time-smearing the power required by an amplifier on transient program peaks.
Or, as Drubin alluded to, some aftermarket cords may be doing something active, like intentionally filtering high frequencies. Or as Sean says, the degree of difference might be directly correlated with a component's power supply quality (or lack thereof). But the bottom line is, if you look at a component's power supply as being in the signal path (which I do, since the output of any amplification device represents modulated wall power, and *not* the original input signal passed straight through), then it only makes sense that purer incoming wall power could result in a higher-fidelity output signal. So in an age when most audiophiles are in some way 'conditioning' their power after the wall socket before it hits their components, it additionally only makes sense that you would strive to preserve that powerline purity between the power conditioner and the gear.
The 'last 3 feet' argument-against also overlooks the real possibility that if you could somehow replace all the powerline wire going back to the breaker box, or to the utility pole - or even all the way back to the power-generation plant - with the same aftermarket PC cabling you use from the wall, then the sound would just get even that much better ; simply because you can't doesn't mean you won't be able to hear *some* (albeit less) potential improvement merely by upgrading the final cord (and particularly if *that* happens to be one of the 'weakest links' in the overall power-delivery chain).
But: I agree that the degree of difference made by aftermarket PC's is probably often oversold to some extent or another, and if you can't hear an improvement for yourself, then just consider yourself lucky (monetarily speaking) and continue to listen happily through your stock cords.
P.S. - It is interesting to note that the brand of amplifiers reputed to have some of the most sophisticated power supply engineering in the audio world, the Halcros, nevertheless are now sold outfitted with base-model Shunyata upgrade PC's as stock. Yes, this could just be clever cross-marketing, but it is highly doubtful that Halcro in any way needed to do this in order to help sell their premium-priced product. Given their 'slide-rule-driven' company image, I suspect they've sincerely determined that anything less would truly be insufficient to serve as their 'stock' cord, kind of a remarkable thing when you stop to consider that at these price levels, many if not most of their customers will be intent on immediately upgrading to ultra-premium cords anyway.
