It is a tempting approach..
Actually the 1.6 is simply the crop.
The optics behind it is as follows: Assuming equal pixel quality btwn both, if the 40D has the same pixel pitch than it would not differ than simply cropping the corresponding inner portion of the 5D frame. Its only the pixel pitch difference that gives you an actual "mag" difference and those hold only to the extent that the modulation transfer function (MTF) on a pixel base is equivalent. The MTF on the entire optical train, including the lens, needs to be considered in this case. So the real "mag" rather than the empty "mag" varies from lens to lens (and within a lens on aperture, zoom, focus distance and even capture wavelength).
Think of it another way. If it were not so, one could make a very high "mag" camera by simply using smaller and smaller chips. Instead of different lenses, manufacturers would simply offer a camera that uses fractions of the detector. You want an effective 24000 mm lens, dial in a 10 by 10 pixel subarray.
The pixel density and pixel quality is actually the issue (as well as optics quality-think of the cheap telescopes or microscopes offering thousands of mags).