For clarity, I think we all agree that shimming is sub-optimal. The method Thom and I described and the link I posted are a band-aid for a problem that shouldn't exist, at least not on pricey tonearms. Shimming works for azimuth but it also has sonic side effects in at least two areas:
MOUNTING RIGIDITY
The more rigid the coupling between cartridge and headshell, the more accurately cantilever movements are translated into electrical signals. Looseness in the cartridge mount slurs transient responses, reduces amplitudes, adds overhang to every note and raises the sound floor - slop, slop, slop.
The solution is self-evident: if you must shim for azimuth, use a rigid, non-compliant material.
ENERGY TRANSMISSION
As Larryi already described.
The precise sonic effects of altering energy transmission between cartridge and headshell will vary with individual components. Nevertheless, inserting two new material interfaces increases the frequencies that will be reflected back into the cartridge. That necessarily raises the sound floor.
Shimming's a reasonable band-aid, that's all.
And what Syntax posted! ;-)