http://dl.dropbox.com/u/10701954/Tax...culations.xlsx
Just made up a spread sheet to calculate this out. Adjust assumptions and scales as desired. I have it set for a build I was trying out for man. O3P1C3G3
Bottom line is that overtaxing at 1.2 in your capital brings in about 1000 gold extra within around 20-25 turns at which point the total difference begins to decrease. At 1.3 its around 1500 gold extra by around turn 20. Your talking about an extra fort or two within the first 20 turns. That's a pretty big deal.
Overall I would say its worth it. Especially in CBM. In vanilla the trade off is a bit more questionable since foresters have a lower patrol and you almost always need 2.
However, as MA Man you also have an incentive to get up a lot of extra forts. You don't quite live and breath the stink of castle walls the same way was LA Man, but you do have some advantages they don't. Namely nature mages coming out you ears to provide supplies, and a bunch of stealthy, high strength wardens (if you're playing CBM) to sneak in and hold up the walls. Given this you can probably have a couple forts crank out a forester or two for 20 gold to patrol some provinces, leading to more gold, leading to more forts, leading to more mages. True if you used a capital recruitment turn to pick up another forester you would lose some research, but ultimately you get more if you have more castles cranking out mothers and bards.
As far as optimizing the strategy, you want as many positive scales as possible. Since the tax rate is multiplicative with any scales, the more positive modifiers you have the more this trade off tilts in favor of over taxing. For instance with O3P3G3 you can get almost 1800 extra gold by turn 21 at 1.3 tax. Take it up to 1.5 and you can get almost 2700 extra by turn 20. That's two extra forts, two labs, and almost half a temple.