I did test it, a province at 120% tax with 3 growth holds its population steady. I didn't patrol down the unrest, so it was just the death from the taxes. Might want to redo your spreadsheet if you're making your predictions based only on the death from patrolling down unrest, as this makes a rather large difference, especially for heavily populated provinces.
Wolfram Alpha has a use! (oh god, edit, it mangles the url when the input is in it)
Input this crude little thing, but it'll do the trick: f(t+1)=(f(t)*c)-b,f(0)=x
Just use the recurrence relation equation at the bottom for any time t to predict the pop with growth/death rate c and constant amount of deaths b starting with initial pop x. Grab the summation or integral from 0 to maxt - 1 to get total income over that time. And the common growth formula for a prov would be the simple e^kt, or initial pop * e^ln(growth/death)t
I might whip something up in maple, but this seems sufficient...