Assuming this was a serious proposal, I believe you will get few takers for this taxcoin, but if you implement it, perhaps I am wrong.  

As I recall, there have been other similar proposals for seizing inactive coins and this is just a slower proposal where instead of seizing them all at once you want to seize them slowly.  This would probably destroy such things as proof of burn, puzzles, etc.  It could also change assumptions and incentives - once an address is close to the “tax period” miners might refuse to mine a transaction with it to collect the tax. Also, from a technical perspective, how is the “fee automatically deducted from an address”?  And I’d also add this could result in a huge number of transactions, enough to potentially fill all available space and result in the inability to have other transactions or even get every “address” active every X days.  There are other issues too. 

Of course, the way to test is to fork your software and see how many people will run it.  :-)