202 Will Regulate
202 Will Regulate
Coming up today: Regulators hit the war path, Bitcoin goes bonkers for NFTs and we play with blocks on Ethereum.
In the meantime, here’s your weekly Bankman Fried Bit: Sam got busted by the Feds for twice using a VPN, a service that allows him to encrypt and hide the sites he is visiting. His lawyers claim that he was merely using the VPN to access his NFL Game Pass account in the Bahamas to watch the Super Bowl and Championship Games. His parents certainly seem like the type that don’t own a TV, but hasn’t the dude ever heard of FoxSports.com? I’m pretty sure you can watch the NFL with a U.S. Internet connection. I mean it’s not like he’s trying to catch Hawthorn vs. Geelong. I’m skeptical.
🟢 The Empire Strikes Back
(Big Idea: Regulation)
If you’re the type of freak out, this was a pretty good week to freak out about crypto regulation as the SEC took two major actions and seems to be gearing up for more. The short version seems to be that regulators did absolutely nothing to stop FTX or any of a dozen other major crypto eff-ups, but now they are newly emboldened to crack down by going after the parties they are actually trying to do some of the right things. And while there is more than enough reason to be concerned, I’ll zag a little here this time and try to walk through some reasons that these recent enforcements might not be quite as horrible as they seem.
Last Thursday, Kraken, one of the biggest U.S. exchanges, settled with the SEC agreeing to end their staking program and pay a $30 million fine. As we’ve discussed, virtually all cryptocurrencies use a system called “Proof of Stake” to secure their network. Users “stake” their coins and risk getting penalized if they behave badly, but in return, help process transactions and receive rewards for supporting the network.
There are a few unique things about staking. First, you can stake without ever giving up control of your assets: you still own them and they are attached to your wallet. It may take a certain period of time to unstake them and be able to move or sell those coins, but they always remain yours.
Second, if you don’t stake, you are going to suffer from inherent inflation. To incentivize users to support the network, those that stake receive newly issued coins. If you stake, this yield helps protect you from this natural inflation, but if you don’t, you will lose out over time.
But staking on your own is technically complicated, computationally expensive and economically prohibitive. On Ethereum, for example, you need a minimum of 32 ETH (roughly $50,000) and the technical ability to run your own node to stake. So, understandably, the market has moved to staking specialists that you can “delegate” your coins to, pay some commission and receive most all of the returns.
Kraken, amongst other exchanges, had set up a way for users to do all of this through the Kraken interface. There are other ways to delegate your coins to staking services if you hold your coins directly, but if your coins are stored on an exchange, your only option is to stake via that exchange..
The SEC, however, deemed the Kraken process a security because investors “lose control of those tokens and take on risks associated with those platforms.” Of course, users already lost control of the tokens as soon as they end up on the exchange (as we found out with FTX) so I’m not sure I get the point. But the “incomplete” disclosures and the aggregation of coins across multiple investors seemed to be enough for the SEC to want to step in.
Then, on Monday, the New York Department of Financial Services ordered Paxos to stop issuing any new Binance USD, a stablecoin pegged to $1 offered in connection with Binance, the largest crypto exchange in the world (by far). The WSJ reported that this came in connection with a Wells notice issued to Paxos alleging BUSD is an unregistered security. There is currently around $15 billion of BUSD outstanding and roughly $700 million BUSD was redeemed in the first day after the announcement.
The whole question of what is and isn't a security has probably been the single biggest issue looming over crypto for the past four years. Coinbase published a blog post on why they don’t think staking counts as a security. And the stablecoin issue seems even more confusing to many since an “expectation of profit” is one of the key tests and stablecoins (at least currently) offer nothing more than the hope that you can get your dollar back when you want it.
People also found the SEC’s methods distressing. Both Kraken and Paxos would broadly be considered “good actors.” Both are U.S. based, highly regulated and have aimed to engage in constructive discussions to create acceptable products. On this front, the SEC has been consistent in its refusal to be helpful and has regulated by penalty, leaving industry players to try to decipher exactly what parts were in violation and which might be OK.
Hester Peirce, the most crypto sophisticated of the SEC commissioners, dissented on the Kraken case and summed it up pretty well:
“A paternalistic and lazy regulator settles on a solution like the one in this settlement: do not initiate a public process to develop a workable registration process that provides valuable information to investors, just shut it down.”
As Warren G sang “Regulators. We regulate any stealin' of his property. We’re damn good too.” [Then again a few verses later he pointed out, “I got a car full of girls and it's goin' real swell. The next stop is the Eastside Motel,” so maybe he wasn’t quite talking about the SEC.] The point still stands however. To a hammer everything is a nail. And the tiger didn’t go crazy, the tiger went tiger. If you were the SEC (or really any agency regulating crypto) and a) you had done nothing to stop FTX (and in fact seemed to be cozying up to them), but b) had the air cover of multiple crypto crashes, this would be a very good time to appear to be doing something significant.
And there may be some extenuating circumstances that make these recent rulings less shocking. With Kraken, yes, staking is subtly, but substantially different and really quite reasonable, but the exact nature of how Kraken was doing it was perhaps a bit too closely similar to things that did, in fact, recently blow up. Celsius promised they would hold your crypto and give you yield in return. Of course, Celsius then basically took that money, bought themselves Lambos, hookers and blow and put it all on red hoping to make it back. Meanwhile, Kraken, by all appearances, seems to have been very responsibly passing on sustainable crypto economics to its customers. But…. the basic premise looks pretty similar despite there being very big differences in how the yield is generated.
As for Paxos, the suspicion is that the real target is Binance. Binance’s market share has grown to roughly 66% of the crypto market since FTX’s implosion and is largely out of U.S. control. BUSD may be one of the vortexes in to Binance and there are hints of this in some of the filings.
More broadly, there is an ongoing question is this is the first salvo or more a sign of things to come. The SEC didn’t attack staking directly, just a synthesized version of it. And they haven’t come after Circle, the issuer of USDC, the cleanest stablecoin we have. This current round seems to have more legal pushback and we will see what the outcome is of the long running Ripple litigation.
Ultimately, these agencies enforce Congress’ rules and without additional Congressional clarity, degrees of enforcement will be dictated by the political environment. I, naively, hope that productive legislation may yet come from Congress in the next two years. The Telecommunications Act of 1996 was passed under Democratic leadership and largely took a hands-off approach to the Internet. Imagine an alternate world where launching a website required getting a government permit and pre-registering what services would be offered.
I doubt the SEC will be able to stop the technological impact of crypto. They can certainly slow its development in the U.S., but that will merely push more and more efforts offshore and completely out of the hands of U.S.’s ability to influence the path it takes.
🟦 You Know What Bitcoin Needs?
(Big Idea: NFT Innovation)
Bitcoin, semi-famously and almost entirely intentionally, doesn’t really do anything. That’s kind of the point, it’s just money (or some version of money or electronic gold or however you choose to define it). And the only real add-on in Bitcoin’s history has been the Lightning Network which basically just sends money faster and cheaper.
When building houses or software products, there’s the old adage, “Fast, cheap or good. Pick two out of three.” Blockchains have a similar “trilemma”: Pick two of Speed, Security or Decentralization. Bitcoin, above all things, has picked Decentralization. I started in Bitcoin in 2017 just as the Blocksize Wars were heating up. Some people wanted to increase the size of Bitcoin blocks to allow for more transactions. Others wanted to keep block sizes small. Smaller blocks make it easier for lots of people to run nodes which therefore increases decentralization. Even today, the entire Bitcoin blockchain, every transaction ever sent, only takes up 450 Gigabytes. You can drop $20 on a 1 TB memory stick on Amazon and have enough room for all of Bitcoin and multiple seasons of Gold Case.
Bitcoin Maxis, those that believe that Bitcoin is the one and true blockchain, have pushed to keep Bitcoin simple and slow to evolve. Its limited functionality reduces surface area for attacks and its slower pace of change improves security. But there are still changes and updates that come along. One of the bigger releases in recent memory was called Taproot and it went live in November 2021. Most of the improvements were around security and privacy and all were highly technical, but the changes also inadvertently opened the door for some new functionality.
Most notably, Taproot makes it cheaper to store arbitrary information within a Bitcoin block while also allowing a single transaction to use up the maximum 4 MB block size (a 50,000x increase over what was previously permitted). That space can be used to store a link, an image, a video or more.
Then, last month, a project called Ordinals launched. Ordinals take the very smallest unit of Bitcoin, the Satoshi (1/100,000,000th of a Bitcoin) and gives them each “individual identities.” This is basically like putting a serial number on a penny.
Now put the two techniques together and you’ve got two great tastes that taste together: NFTs on Bitcoin. The ordinals turn something that is normally fungible into something unique by assigning a specific identity to a specific Satoshi. Then, that Satoshi can be “inscribed” with some specific piece of content.
It took about a month and a half for people to figure it all out, but in the last two weeks, Ordinals have taken off. And I’m sure you can imagine where this is going. The first ordinal seems to have been a pixelated Mexican skull, but, obviously, the second was a Crypto Dick Butt. [I heard that game designers have a metric called TTP or “time to peepee (edited for spam filters)” which measures how long it takes for players in a game to generate genitalia. Ordinals named that tune in two notes!] Since then, almost 100,000 ordinals have been inscribed: mostly images including Crypto Punks and Bored Apes (the two biggest ETH based NFT collections), but also love notes and one piece of text that just says, “Basil!” . You can scan the most recent or see them all here and there’s a good dashboard of all of the on-chain measures (counts, fees, etc.)
Bitcoin Maxis are not amused. “inscriptoooors are spray painted names on the porous marble of a cathedral.” But they’re also a bit stuck since the whole idea of Bitcoin is to be decentralized and censorship resistant. There’s tons more to cover on this including the implications for Bitcoin’s fee model and security budget. But Ordinals are a good example of what I find exciting about crypto in general: we really just don’t know where the heck all of this is going to take us. That’s super scary to a lot of people, especially if you work for virtually any three initial government agency, but it’s incredibly powerful and so very much like the early days of the Internet.
◆ Building Building Blocks, Part 2
(Big Idea: How Blockchains Work)
Last issue, we started discussing the process of putting together transactions to fill a block on the blockchain. If you didn’t read it last week, you can catch up on it here. But to quickly recap, the general idea is that when miners choose what transactions to include in a block, there are degrees of flexibility. Normally, this is driven by the fees senders are willing to pay, but there are always exceptions.
We introduced these ideas using Bitcoin because of its relative simplicity, but today we’ll move on to Ethereum which has much more room for shenanigans within a given block. The main difference is that with Bitcoin essentially all you can do is move Bitcoin from one address to another (as of one month ago, that is). Ethereum allows you to do things, even multiple things within a single block, so the order in which transactions happen matters intensely.
The simplest example of this is what’s called a “flash loan.” In a flash loan, a user can borrow assets without any collateral so long as the loan is paid back in the same block. If the block doesn’t end in the loan being paid back, the original loan is canceled. Why would you want a flash loan? Turns out there’s lot of reasons, most of them moderately malignant, but arbitrage is one of the simplest:
Trader Joe (not his real name) borrows 1,000 ETH (~$1.6 million) from Aave, a decentralized automated lender.
TJ immediately sells that ETH for USDC at $1,620 on Uniswap, an automated market maker (AMM).
TJ buys back that ETH at $1,600 on Sushiswap, a different AMM that may has different pricing. TJ now has a paper profit of $20,000.
TJ pays back the original loan. He pays Aave 0.09% for the loan and also has to pay gas fees to the ETH network to process the transactions. The loan cost about $5,000 and the gas fee is maybe $20 (it depends) so TJ made ~$15,000 on this theoretical transaction.
Obviously, the order of these transactions is important since none of it works out of step. And it’s also critical that it all happens in one block because if it doesn’t, there are much larger amounts of risk introduced on all sides.
Vanilla arbitrage is no big deal and helps keep markets aligned, but this ordering of transactions creates other exploitable opportunities that are broadly negative for the environment. One example of this is a “sandwich attack”. If you recall from last week, pending ETH transactions get submitted via a local node to the global “mempool” where they are visible for everyone to see and for miners to include in an upcoming block. In a sandwich attack, bots search for large trades that will move prices and then front run those trades, resulting in worse execution for the original party. Here’s how it works:
Wanda Wegman (also not her real name) submits a transaction to buy 1,000 ETH with USDC on Uniswap. This is a simple trade so she only needs to offer $25 in “gas” (transaction fees) to get the trade processed.
Billy the Bot (real name) sees this transaction in the mempool. He knows that Wanda’s order is big enough to move price the price of ETH on the AMM and he can make it move even more. Billy submits his order, but he offers $50 in gas to make sure that his transaction gets executed first. He buys those 1,000 ETH at $1,620 and then the price moves up to $1,650.
Wanda buys her 1,000 ETH, but has to pay $1,650 per ETH instead of $1,620, costing her $30,000. Her purchase pushes the price further up to $1,680.
Billy then backruns the trade, selling his 1,000 ETH at $1,680 netting a profit of $60,000 absent any other fees.
Here’s a real block that shows this all at work. Let’s break it down:
This ETH block included 144 total transactions. On average, the transactions were waiting 12 seconds in the mempool before being included in this block. Remember that on Ethereum, there are new blocks every 12 seconds so this makes sense. If you look at the details, you can see that one transaction had been waiting for 1 minute and 54 seconds.
But let’s dive into the sandwich. Some user, 0x4A5Bb, was trading 25 ETH for 13,921 LDO (a token of a liquid staking provider called Lido that we may spend more time on soon).They submitted their transaction 3 seconds before the block was created (the “delay” column shows how long the user waited). A user literally called “jaredfromsubway.eth” jumped in and front ran our first user. Jared started with 7.55 ETH and traded it for LDO. Then after 0x4A5Bb’s trade went through, he bought back his ETH and ended up with 7.63 ETH, a profit of $183.78.
I know I’ve lost most of you by this point, but I can tell a few of you are already jumping ahead. Why does the “delay” field for the frontrun and the backrun say “miner”? Well, that’s because it’s the miner that is doing the sandwich attack. That’s right, the call is coming from inside the house!
A reasonable way to think about this is that the miner is just hanging out, picking which transactions to include in the upcoming block and then what order to put them in. The website I’ve pointed to even shows some options of how that specific block could have been ordered. “Fair order” is FIFO, who’s been waiting the longest. “Gas order” is you order the transactions based on who paid the most in gas fees. Then “Block order” is how the transactions were actually ordered.
But instead, the miner noticed that there was a tasty morsel that they could frontrun. This is how we come back to MEV: maximum/miner extractable value. Instead of leaving that profit for some bot to come along and snap up, the miner took it themselves.
That’s enough playing with blocks for today, but make sure to come back next time for the exciting conclusion where we’ll discuss Flashbots and some of the other ways to “solve” MEV.
This Week's Freezing Cold Take
Good to see regulators are consistent over eras:
“One provision of Massachusetts securities law stipulates that per-share offering price can’t exceed 20 times earnings. Apple last year earned 24 cents a share and the $22 offering price is about 90 times earnings.” [Editor's note: HAHAHAHAHAHAHA!!!]
As always, thanks for reading. Send me questions and please share with your crypto curious friends.