Ok here's my take:
Due to quantity and/or quality of glass in a lens, there may be a slight difference in the amount of light reaching the image sensor. This may be more in better vs. worse lens, or may be less(!) HOWEVER, this difference will be minor, in the order of <1/3rd stop.
I did the following test with two of my lenses. The 1st photo was taken with an EF 28-90mm f/4-5.6 III lens (The kit lens from my EOS 300X 35mm camera. Image quality is probably worse than 18-55mm kit lens). The 2nd photo was taken with my 24-105mm f/4 L IS USM lens, which by general consensus is a pretty good lens.
Camera set to manual, zoomed in max to 90mm, metered using the built in meter to f/5.6, 1/60 at ISO100.
Next, change lenses, zoomed to approx same image size using Live view, kept metering settings exactly the same.
Imported into Lightroom, white balance set to "Daylight", exported as 80% sRGB JPG.
Ignoring the obvious quality differences, the image from the better lens is DARKER!!!! In other words, the better lens (the L lens) let LESS light reach the sensor. There goes the OP's buddy's theory
As far as I can tell, the difference is roughly 1/6th stop.
According to my camera's meter, the 2nd photo using the L lens was underexposed (remember I metered using the first lens) by 1/3rd stop. Obviously the camera can't indicate underexposure by less than 1/3rd stop. Secondly you may have noticed I bumped my tripod and the two photos aren't exactly aligned, so that may also affect the meter a bit.
What does everyone make of this?? The only reason I can think of other than the differing amounts of glass, is that due to the quality of the light reaching the light meter being better in the case of the L lens, the camera is able to meter more accurately?
If I have some time I'll upload these to flickr for everyone's perusal.