I'd think there is one situation where using at least a *longer* extension, if not stacked extensions, could reduce effectiveness, if not torque. If you used one, or a combination of, extensions that had a significant amount of give in the metallurgy (think spring steel - like a torsion bar), you could turn all you like (wasting effort in the process), the bar would absorb your effort as potential energy and, until you reached a certain point where enough energy was transferred to the fastener to break it loose, it would simply release all that energy back to the ratchet if you let go.
Not torque, but work, put into the system and stored as potential energy, and not being transferred to the fastener.
The other idea, though it's really insignificant, is that the mass of more extensions (greater than the mass of a single extension of the same total length) has a greater moment of inertia to be overcome, which would be considered a parasitic loss, like if you replaced a carbon fiber driveshaft with one made of cast iron. This is really an insignificant difference for socket extensions, though.
This is all moot, no stack of extensions that could still be reasonably put to use would have that much give before transferring the energy to the fastener. Even if the chrome vanadium steel could absorb the energy, it would crack at the square drive edges long before twisting. Most of us have probably been witness to this more than once. Which is why I'd say you wouldn't want to stack extensions due to the increase in points of failure. If you had to hoist a load by a chain, would you rather have to worry about two links to have fail or two hundred? A chain is only as strong as it's weakest link.
I think Sunex makes some 2- and 3-foot impact extensions in 3/8 and 1/2"-drive for transmission work. Then again it might be Grey Pneumatic, but I remember seeing them and thinking I'll probably wish I had those oddball pieces someday.
