The signal in Marathon virtue-signaling

Last week the Marathon mining pool generated plenty of controversy by carrying through on a promise to mine “clean” blocks. First the adjective is misleading. Given renewed focus on the energy consumption associated with Bitcoin, it would be natural to assume some environmental connotation, specifically using renewable sources to supply the massive amount of electricity required for producing that block. Instead for Marathon the measure of block hygiene turns out to involve an altogether different yardstick: compliance with OFAC sanctioned addresses list. The Office of Foreign Assets Control is part of the US Treasury Department. It is responsible for maintaining lists of foreign individuals that US companies are barred from doing business with due to national security and foreign policy concerns. In other words, OFAC is the reason that El Chapo or GRU officers can not sign up for a credit card, open a savings account or apply for a mortgage with a regulated US bank.

OFAC has long taken an interest in cryptocurrencies. It has even sanctioned blockchain addresses on Bitcoin, Ethereum, even privacy-friendly coins such as Monero and ZCash. Regulated US-based cryptocurrency companies already take these lists into account. For example they can block outbound transfers by their own customers (classic scenario is stopping payments to ransomware operators) or freeze incoming transfers from black-balled addresses to prevent those from being laundered by trading into another asset. In that sense, there is nothing new about some blockchain addresses becoming “radioactive” in the eyes of financial institutions. Where Marathon has crossed into uncharted territory is applying these rules to mining new blocks, in a way that affects all participants globally. In the process, it opens a can of worms about the concentration of mining power and whether governments can exert influence on an allegedly decentralized system by squeezing a handful of key participants.

Meaningless gestures

First let’s start with MARA pool. With an estimated share of less than 8% of total hashrate, this clean mining campaign is unlikely to make a dent. (Incidentally that estimate comes from a January Marathon press-release, as an upper-bound on the hashrate achievable if 100% of its capacity were directed at bitcoin.) When Marathon wins the race to mint the next block— which happens 8% of the time on average— it may exclude a pending transaction that would otherwise have been eligible according to standard rules. But other miners remain unencumbered by that self-imposed rule and are happy to include the same TX next time around when they win the race for the minting a block. Crucial point is that once any miner includes the transaction, Marathon is stuck mining on top of that block. Quite ironically Marathon is helping the sanctioned actor now, adding more confirmations to the verboten transaction and helping push it deeper into blockchain history. The net effect from the sender perspective is a slight delay. On average, sanctioned address will find their transactions taking slightly longer to confirm than other addresses. At 8% of share, the difference is imperceptible. Even at 20% it would barely register. Recall that only the initial appearance on block can be delayed by Marathon “clean mining” policy. Once included in any block by any other miner, additional confirmations will arrive at the regular rate. Considering that most counterparts require 3-6 confirmation, the additional delay on the first block is negligible. This is a scenario where the intransigent minority can not enforce its rules on the majority.

Corollary: Marathon “clean mining” is pure virtue signaling.

A conspiracy of pools?

What if additional pools jump on the clean mining bandwagon? Imagine a world where miners are divided into two camps. Unconstrained miners simply follow consensus rules and profit maximization when selecting which transactions to include in a block.  Compliant miners on the other hand observe additional restrictions from OFAC or other regulatory regimes that result in the exclusion of certain transactions. At first this endeavor looks a doomed enterprise regardless of the hash-rate commanded by the compliant miners. As long as any miner anywhere on earth is willing to include a transaction— including the proverbial lone hobbyist in his/her basement— it will eventually get confirmed. At best they can slow down the initial appearance. That is still problematic, in that it breaks the rule of fungibility: one bitcoin is no longer identical to another bitcoin. Some addresses are discriminated against when it comes to transaction speed. Still this looks like a minor nuisance, considering that funds will eventually move. Worst case scenario, a sufficiently motivated actor can temporarily rent hash-power to mine their own censored transactions. That suggests any attempt at extending sanctioned address lists to mining is tilting at wind-mills unless it can achieve 100% coverage.

In fact total censorship can be achieved without control over all miners. Once the share of compliant miners approaches the 1/2 mark, the game dynamics shift. In effect compliant miners can execute a 51% attack against the minority to permanently exclude transactions. As before consider the scenario where a miner who is unaware of or deliberately running afoul of OFAC rules mines a block including a transaction from a sanctioned address. Compliant miners now have an option that is not available to Marathon: ignore that block and continue mining on a private fork without the undesirable TX. Unconstrained miners may have a head-start after having found the block first. But the majority will eventually catch-up and become the longest chain, resulting in a chain reorganization that erases all traces of the censored transaction from bitcoin history.

Game-theory of righteous forks

Would regulated miners invoke that nuclear option? Initiating a fork to undo some disfavored transaction is an expensive proposition— even for the side guaranteed to win that fork battle. It is disruptive for the ecosystem too: consider that on both chains, block arrival times will slow down. Instead of 100% of hash-rate being applied to mining one chain, it is split between two forks, but with each fork having the difficulty level of the original chain. (Since these forks are expected to be resolved after a handful of blocks, difficulty adjustment will not arrive in time to compensate for the reduced hash-rate.) On the other hand, compliant miners may have no choice. Their actions are not driven by profit maximization alone; otherwise they would not be excluding perfectly valid transactions in the first place.

In terms of financial incentives, there is a silver-lining to being on the winning side of the fork: block rewards previously claimed by unconstrained miners are up for grabs again, as alternative version of those blocks are produced. Since block chain history is being revised, the coin-base rewards can be reclaimed by compliant miners who supplied an alternative version of blockchain history on the winning side. When history is rewritten by the victors, coin-base rewards belong to those responsible for the revisionism. This incentive alone could compensate the regulated majority for going through the trouble of having to initiate 51% attacks in the name of keeping the blockchain clean. On the other hand there will be unpredictable second-order effects, since miner behavior is visible to all observers. Heavy-handed chain reorganizations and censorship may result in a loss of confidence in the network and corresponding depreciation of Bitcoin, hurting the miner bottom line again.

Unconstrained miners in the minority have even more to lose: not only will they waste resources mining on a doomed chain until the reorg, but they will also lose previously earned rewards for blocks that were replaced in the fork. A rational miner will seek to avoid that outcome. Going along with the majority is the path of least resistance, even if the miner has no ideological affiliation for or against the regulatory scheme in question. As long as the majority is committed to forking the block-chain at all costs in order to avoid running afoul of applicable regulations— the metaphorical gun to the head— game-theory predicts the minority will fall in line. There may be a handful of 51% attacks waged initially if unconstrained miners seek to test the resolve of the regulated ones. Ending up on the wrong side of such a fork will have the effect of quickly reseting the expectations of miners in the minority.

The tipping point may well arrive before 50%. Strategies such as selfish-mining allow smaller concentrations of hash power to attempt to hijack the chain. Compliant miners may even be compelled by regulation into temporarily withdrawing their hash power or even attempting a Hail Mary reorg for a limited number of blocks, before giving up and resuming work on the longest chain even if contains a tainted transaction. While the minority is likely to lose most of these uphill battles, even a small chance of victory and associated redistribution of coinbase rewards could motivate unconstrained miners to play it safe.

Choosing the regulators

What does this portend for the future of cryptocurrency regulation? Marathon may have voluntarily indulged in this bit of meaningless virtue-signaling act, but it is a safe bet that regulators elsewhere are taking note. The concern is not about OFAC or the selection criteria used by the US Treasury for its sanctions. There is a very good argument to be made that ransomware operators, Russian election-meddling groups, ISIS terrorists or North Korean dictators should be cut off from every financial system, including those based on blockchains. Companies operating in that ecosystem in any capacity— exchange, custodian or miner— have a part to play in implementing those policies.

The problem is US regulators are not the only ones drawing up lists of personae non gratae and declaring that transactions by those actors must forever be consigned to the memory pool. In fact the concentration of mining power in China— dramatically illustrated by the drop in hash-rate during a recent power outage— hints at a darker possibility: CCP could point to the Marathon example to strong-arm mining pools into blacklisting addresses belonging to political dissidents, human-rights organizations, Uighur communities and Tibetan activists. Marathon was not under the gun and acted voluntarily. Mining pools in China may have a very literal gun pointed at their head while instructed to comply with censorship rules.

CP