After looking at the timer stuff in perxteak.per, I figured that they are both having different internal clocks. I had only tested timer 1 yet, but when testing both timers...
Timer 0 is 134MHz (called "internal clock with no wait-states")
Timer 1 is 107MHz (called "internal clock with wait-states")
No idea where those wait-states come from, and if they are constant. It might be something else than the Z0/Z1/Z2/Z3 area wait-states.
Another difference (not mentioned in the .per file) is that Timer 1 seems to have the MU bit always set (the bit is readonly).