I expects folks to recoil at the idea of tainted coins, because it's based on breaking fungibility, but if the regulators get fed up they could use this. I think they eventually will.

OFAC currently has a very long list of sanctioned addresses. I have no idea how intelligently constructed this list is, nor how it is updated. As it's trivial to create a transaction spending your bitcoin to yourself but with a different address - to get Bitcoin off this list simply requires one miner to mine such a transaction. It's been convincingly argued that OFAC is basically helpless here (see for example recent discussion BPI paper by Josh Hendrickson and Craig Warmke) : They can supplicate US miners not to process these, but all it takes is one miner, somewhere in the world, to process the transaction, and the transaction is gone and replaced by a new one. There's about 144 blocks mined per day; US miners can't control all of them.

Bitcoin has a UTXO structure, meaning every transaction has at least one parent which it can be traced back through. So if a hacker, (or a paid assassin, pick your least lovable criminal) has a large amount of Bitcoin, we could in theory follow where all that goes forward in time, and attach a taint to it. All this would require from OFAC is software monitoring the blockchain and then adding all descendent transactions to the list. But then what?

Here's a couple easy things. First, legislate that ETFs, DATs, exchanges, or any regulated corporations are forbidden to hold such coins. If these corporations are sent one, they are required to not credit whoever sent the coins, and turn the coins over to the USG (where they are now declared clean, in case they are to be auctioned off, or just held in the SBR.)

Then, to make this much more powerful: extend the taint to the coinbase transactions of blocks that include any tainted transaction. Thus miners who choose to include tainted transactions are tainting all of their rewards. If they want to send their mined Bitcoin to an exchange in the US, they can't. They can send it to an exchange somewhere else, which may or not accept it - if it gets mixed in with their coins, their customers might not want it. If ETFs are a major buyer of bitcoins, by choosing to accept or mine a tainted transaction, you are restricting the future market of that coin to a significantly smaller and actively less liquid subset. Fewer bids and more asks means a lower price when you want to pay your bills.

Exchanges (those overseas) in turn, to stay relevant, will separate the coins. Two markets. Two prices. One for illicit use, the other for ETFs, DATs, would-be Saylors and everyone who's just LARPing about being on Bitcoin standard.

Note that we don't need to forbid people from using these or holding them, that would be hard to enforce and perhaps overly aggressive; just lean on the compliant corporations.

The rollout would take some moderate care, but not too much: After a certain block, say block 1,000,000, there's a list of tainted addresses, and this list is updated continuously. An API provided.

The game theory tracks backwards now. It's clear that a miner situated in the US who may want to sell their block rewards to pay bills won't include transactions that are sanctioned. Other miners have a decision to make: maybe there's a large fee in there. But adding the large fee is an additional cost to using the UTXO - everyone downstream will have to pay a large fee to get their transaction included, making the proceeds less desirable.

Even worse for crooks: If you meet someone in a Walmart parking lot to buy a bike with Bitcoin, even attaching a large fee to the transaction does not mean the transaction will be mined. You may have to sit there with your new Craiglist bestie waiting for hours to get a single confirmation. (Obviously you want to be careful here - next week when he gets busted, authorities will check your Bitcoin address against a list of KYC'd ones and you might have to give the bike back. Despite the fact a coin might be dirty, they still might be able to track it! )

Yes, this creates a hassle for wallets and end users. (This is a wallet and end users problem, not a USG problem.) Wallet providers ping the API as new UTXOs are added (this isn't a hard lift) and flag the tainted ones as tainted. The wallets would also be careful not to mix the tainted with the untainted (In particular, you can't "dust" someone with dirty Bitcoin unless their wallet chooses to mix, the UTXO structure gives these by default separate lineages.)

If the wallet software chooses not to do this, you accept Bitcoin at your own risk.

This would provide a massive disincentive to do things like CoinJoin: It's almost certain someone else in the CoinJoin was dirty, so now you all are. If you really want dirty but un-KYC bitcoin you can have that, but there's still no guarantee.

Now if this the taint is for temporary sanctions purposes, you could always un-taint when the sanctioned party is back in good graces - but the coinbase taint would be forever.

It's clear the same approach could apply to Zcash; taint all shielded address, taint all proceeds from a shielded address.