I doubt that anyone making that claim has gone to the trouble of doing comparisons with and without the covers that were extensive enough and thorough enough to rule out the possibility that the results were muddled by extraneous variables. Examples of possible extraneous variables being things like unequal warmup times, changes in AC line voltage or noise conditions, ongoing breakin or aging of system components, etc.
IMO, particularly in the case of effects such as this one that would seem to be technically implausible and/or quantitatively insignificant, it behooves those making such claims to be especially thorough in their test methodology. That would include, as a minimum, going back and forth several times between having the cord enclosed in plastic and not having it enclosed.
I doubt that anyone has performed such a set of experiments, and even if they did I see no reason to expect their results to be applicable to components whose designs are different than the ones they used.
Personally, I suspect it is nonsense. What I can see making a difference, though, would be having power cords for more than one component in close proximity to each other within a cover, along a significant run length. I wouldn't doubt that that could result in electrical noise that is generated by one component, and fed back into its power cord, coupling into the adjacent power cord of another component and affecting its performance. Especially if a digital component or a power amplifier is involved.
Regards,
-- Al
