The title of this column is a direct quote from Don Wilson, founder and CEO of the large trading firm that bears his initials DRW, at a recent Blockworks conference. He is also on record with thinking markets would be better if all orders were required to be routed to "lit" order books.
The entire Ethereum ecosystem is built around Maximal Extractable Value (MEV) formerly Miner Extractable Value. This is a sort of value taken out of pending transactions. If that sounds a bit like illegal front-running you should go listen to the linked interview because Wilson agrees with you. And in addition to Ethereum a similar dynamic exists in most blockchain ecosystems.
MEV is one form of value extraction when order run through a process where at least some steps are not "lit" or completely open and transparent. Yes, everyone can see the transactions that yield MEV in public blockchain data. But the order prioritization and transaction shuffling and other vagaries of assembling requests to transact into actual settle on-chain blocks all involve non-transparent processes. You can see what sits in a mempool but you cannot see what sits in all possible mempools out there or inside all the machines reading from mempools to assemble them into blocks. Not everything here is "lit."
There is an obvious conflict, and perhaps a problem, when a central feature of many blockchain ecosystems looks like an illegal practice to at least one influential and experienced trader that happens to own a major trading firm. Maybe everything really is fine. But there is at least a chance of trouble here. So we cannot just dismiss this comment as one man's opinion and not important.
And we in fact agree there is more than just a kernel of reasonableness in Wilson's statements. Recall that markets exist for a reason. Financial markets are not a game. They are a tool with social and economic purposes. When the work well there are positive externalities for the world at large. And when they work poorly there can be negative externalities too.
The Purpose of Markets and a Bit of History
The whole reason we have financial markets, and the whole reason governments care about having orderly and well-regulated financial markets, is that they have proven to be an efficient system for allocating resources over the past few centuries. This does not mean markets are perfect in general. And it certainly does not mean any particular market architecture is perfect today. But, overall, markets have done a pretty good job in delivering growth to the world over the past few thousand years.
Japan has had a rice futures market for about 300 years. The Amsterdam Stock Exchange goes back about 400 years. Agricultural futures were actively traded in Greece two or perhaps even three thousand years ago.
Bear in mind that while medieval, even ancient, Japan and Europe may have had laws and legal systems that roughly approximated today we can see commonalities in finance back much further.
Hammurabi's Code from nearly four thousand years ago regulates debt contracts, mortgages, partnerships, and even structures that look a lot like banks and custodians and insurance in ways we would recognize today. At the same time Hammurabi had a sort of consumer protection regulation requiring carpenters to replace any of a client's slaves they killed while doing construction work. That clearly comes from a society with markedly different views on individual rights and freedoms and such than the modern world. And yet the financial regulatory lineage is immediately recognizable.
Over time we have refined the rules to make markets better at allocating resources. And also to align with the evolving values of society as a whole. The long, old, common thread here is that markets exist to support society and society as a whole has an interest in well-functioning markets.
Complexity for Complexity's Sake
As society gets something from well-functioning markets society should expect to pay something for them. If a market design requires active participants – market makers, agents, banks, whoever – then they need to be allowed to make a profit so the market runs. Expecting parts of the capitalist infrastructure to run on charity is unreasonable. But there are a huge number of ways to organize and fund market activities.
In Scotland most real estate transacts at public auctions. Property agents and variable fees and price negotiations are really not a thing. You show up at the auction, you bid, and someone wins and pays the auction fees and that is it. Yes there are a few more minor details but in broad strokes that is how Scottish property works. To anyone familiar with transacting real estate outside of Scotland this may look ridiculous. It may also look wonderful.
Is that auction-based process better than other systems? Should the government fix auction fees? Should the government itself run the auctions for free and finance operations out of taxes? There are many designs of this one design. And those are all political questions to be resolved in Scotland. But someone has to pay to run the system and costs need to get allocated somehow. If the Scottish setup looks different to you think how many more schemes you have never even considered may be in use already today somewhere. All of which can be arranged with myriad ways for someone to pay for their operation.
It is the same for financial markets. Nobody advocates for governments running all the exchanges, fixing fees at zero, banning market makers and then funding the operations out of taxes. Few even advocate for fixed and rigid requirements on structure to being with. Exchanges compete on speed and price and execution quality and order types and various other features. And markets are full of processional players that add liquidity as their business. Many variants exist and some products – most often government bonds – are actually sold at scheduled periodic auctions.
But these systems exist, and compete, in the service of better markets. Governments sell debt, and wireless spectrum and other regulated assets, at auctions because there are strong reasons to believe they are the best tool for those uses cases . Many market systems exist and they are often adapted to the needs of the economic problems they are intended to solve.
None of these systems is used because complexity itself is good. And certainly none of them is used because some engineers think a weird workaround for a problem they themselves created is clever and should be let loose on the world.
Markets exist to serve society as a whole.
What Are We Really Doing Here?
So what about MEV? MEV is a side effect of the way Ethereum assembles blocks. This problem was first observed in Bitcoin where blocks are also assembled via some multi-party decentralized process.
Bitcoin blocks are assembled via a competitive process. And importantly one does not know who will assemble the next block – or the next 10 blocks – at any time. And simultaneously block space is a scarce resource. So someone that wants their transaction executed "soon" needs to express this view to a number of potential assemblers of the next block and provide them a sufficient incentive to fit a transaction into the limited available space.
This complexity exists as a consequence of using a certain type of decentralized consensus mechanism. It does not exist because it serves market participants outright. It exists because of the way the market operates in much the same way someone needs to pay a Scottish real estate auctioneer's fees but there are no auctioneers fees on the London Stock Exchange or for property transacted via agents in England. Different systems begat different ways of paying for those systems.
Bitcoin has had this dynamic for 15 years. Ethereum is a more complex system with more complex transactions and has had a number of variants of this problem over the last decade.
Years ago people in Ethereum wanted to eliminate MEV and also called it a significant issue with grave consequences . In the early days of Ethereum and MEV the idea naive user transactions would be routinely front-run was viewed as a bad thing. Galaxy Research, source of the "grave consequences" comment, wrote extensively in that same piece about "frontrunning" and lamented that all the possible approaches are "riddled with tradeoffs." Sure. What matters is whether those tradeoffs are worth it for markets and for society as a whole.
Flashbot's MEV-Boost is a complex thing that takes one approach to dealing with MEV. The whole Ethereum consensus mechanism and the proposer-builder separation thing are complex ways of at least accommodating it. So too are Solana's consensus mechanism and that of every other real, used, blockchain. Are these interesting algorithms for engineers? Are they "fun" software to build and run? Sure maybe. But financial markets are not a game and they are not most certainly not a homework assignment.
Markets exist to serve society by efficiently allocating capital. Efficient market designs are often non-trivial and some participants are required and they need to get paid somehow to operate bits of the design. But the design is not the product. Price discovery and resource allocation is the product.
Step back and look at how government bond auctions look and compare it to Ethereum's block processes and MEV-boost. The former looks like it was designed entirely to solve an economic problem. The latter looks like at least some of it was designed to entertain engineers and give them an opportunity to debate "clever" solutions to problems they themselves introduced for no good economic reason.
A reasonable person might look at all the infrastructure around MEV, and all the money being extracted, and say two things. First, they could acknowledge that each incremental bit of complexity is clever and interesting and does some reasonable job of solving some other incremental problem. But then they can step back and note that when all this machinery is needed to mitigate a consequence of a basic design decision maybe that design decision is not worth what it costs overall. Markets are not supposed to be designed to entertain the engineers writing the software that powers the markets.
Decentralized?
All this mess comes out as a consequence of the decentralized nature of the underlying blockchains being used to transact.
But these days the vast majority of the economic activity on those blockchains concerns centralized stablecoins with freeze functions and regulated operators.
If MEV is the price to pay for decentralization and you find yourself trading USDC and still paying MEV what are you really getting in exchange for being front-run?
In practice the answer is that, to pick on USDC here for a minute, Circle will not actually go out and censor your transaction unless you are some kind of criminal. Put aside for a moment the whole "censorship resistant for who?" question and just look at a recent incident in which Circle froze 16 addresses many of which were openly linked to completely different non-criminal services . Hot wallets of service providers were frozen for reasons that remain a bit unclear.
If Circle is going to freeze people seemingly-randomly then you are not getting much for your MEV are you? Yes, we acknowledge these freezes were not random in the sense that Circle picked wallets at random and froze them. Some legal process existed and Circle followed that. But from the perspective of any users of these services they sure were random!
The services themselves with the frozen funds are not implicated in a crime here – they are merely third parties holding funds linked to some case commingled with a large quantum of other user's funds. As an "other user" this does look random. The existence of a legal process is little comfort to some "other user" with funds frozen on a platform because it sits adjacent to something being investigated.
So now we are ready to tighten up Don Wilson's claim a bit. And while sharpening the scope we will also sharpen the sentiment. MEV is dumb and wasteful when the transactions at issue pass through a central gatekeeper anyway. Users that pay MEV when transacting non-bearer assets are basically being robbed by market designers. This is true at least in a basic moral sense because there is no good reason for MEV to exist in centralized systems as you can achieve the same properties in MEV-free designs.
Anyone advocating for systems built around MEV to transact primarily in centralized stablecoins is advocating for users to get mugged. We expect most of the people in favour of these designs either benefit directly from MEV, like MEV because they enjoy playing around with the software or just like creating more infrastructure and process and work for its own (investment) sake. Oh, and a large measure of these people probably still do not know that substantially all the stablecoins out there have freeze and seize functions such that they are fully centralized.
Don Wilson is right. And ignorance of why is not an excuse to disagree with him.
