At first glance, one is tempted to say, So what? But wait! Who challenges an SEC subpoena all the way to the second highest court in the land? Call me cynical, but do you think, maybe, someone who has something to hide?
The CEO in question was Do Kwon, of Terraform Labs. Terraform is a Singapore entity and Kwon is South Korean. In late 2020, while the rest of us were hunkered down to avoid COVID, Terraform introduced the Mirror Protocol, which is what the SEC wants to know about. In the online White Paper, Terraform stated:
See https://docs.mirror.finance/. So mAssets allow you to profit from assets without the hassle of owning them, which is like eating your cake without having it at all. The mAssets contemplated are mirrors of US-traded stocks. For example, you can find quotations for mAMZN, mDIS, mGOOGL, mGS, mMSFT, and mSBUX, among others, which purport to mirror shares of Amazon, Disney, Google, Goldman Sach, Microsoft and Starbucks. See https://mirrorprotocol.app/#/trade.
On the one hand, you can argue that if it trades and quotes like a stock, then it's a stock. Except that it's really not. Unlike, say, an ETF or a mutual fund, there's no pool of Amazon shares backing up mAMZN. mAMZN is at best a shadow stock, or a phantom stock, like those that many securities firms like to give their employees to make them feel they belong. Arguably, mAssets don't pass the Howey test, in that since they are not passive interests in any profit making enterprise, but are simply figments of the imagination of who own them, they are not even securities.
Terraform did argue that mAssets were not subject to SEC jurisdiction, but the court side-stepped that argument with a back-of-the-hand footnote. It's too obvious why the SEC might be interested in the Mirror Protocol -- and also why Terraform might not want to answer questions about it. Arguing they are not securities just begs the question the SEC is investigating. Instead, the court focused on Terraform's two technical arguments against service of the subpoena, neither of which needed a 10-page opinion. See https://assets.law360news.com/1500000/1500866/terra%202nd%20circ.pdf.
The first technical argument was based on an SEC rule that requires subpoenas to be served on counsel when a party is represented. The SEC had been in contact with Terraform's attorneys for weeks. Sure, except that Terraform had not authorized its counsel to accept service, and even at oral argument Terraform's lawyer would not confirm that he was so authorized. As the court said, Terraform's argument would "produce absurd results by allowing a party to insist on service through counsel, but allow the party to block said service by not authorizing their counsel to receive any filings." It was just the sort of argument the overly-STEMed would concoct, steeped in a logic inversely proportional to its common sense.
Terraform's second argument was that it didn't do any substantial business in the US. That one violated the gag rule-in that it's a tough one to make without gagging. After all, mAssets are bought and sold in the US. The stocks that mAssets mirror are US companies traded on US stock markets. And Do Kwon himself was served in New York City while attending Mainnet 2021. See https://blockworks.co/sec-subpoenaed-terra-founder-do-kwon-at-mainnet-2021/. Mainnet bills itself as "An immersive, agenda-setting annual summit . . . gathering crypto leaders/builders/investors/
operators." https://messari.io/mainnet. Clearly, Kwon was in the world's largest financial market to promote Terraform's cryptofinance products, especially, at that time, the Mirror Protocol.
The SEC served its subpoena in September 2021, when it and rest of us didn't really know much about the Mirror Protocol. But we knew and know this much-the Mirror Protocol is based on algorithms that are designed to recreate in mAssets the economic attributes of real assets. How this actually happens is shrouded in some mystery, but later events gave us a clue.
You see, Terraform is also the company behind TerraUSD, the cryptocurrency that was supposed to be a stablecoin pegged directly to the value of the US dollar. In mid-May of this year, TerraUSD crashed, so that now instead of being worth $1, it is worth less than a quarter.
That wasn't supposed to happen, and that it did is a cautionary tale about algorithmic protocols and crypto in general. TerraUSD wasn't actually backed by dollars, but by a reserve of bitcoin. In theory, whenever the value of TerraUSD dropped below the value of a dollar, the algorithm (using an intermediary coin called the Luna) would spend its bitcoin to bring the value of TerraUSD back up to a dollar.
All of this highlights a fact -- when virtual reality collides with real-world reality, the real world always wins. It's a lesson everyone learns sooner or later, and the crypto bros are learning it now.
That's not to say there's no place for crypto-technology in the financial markets. A host of inefficiencies can be eliminated by representing corporate equities as cryptosecurities maintained on some form of blockchain, rather than analogs of paper certificates on ledgers controlled by transfer agents, clearinghouses, and brokerage firms. But then cryptosecurities would be real securities, and not mere avatars like mAssets.
mAssets will never truly mirror real assets because they will always be riskier, and in ways we won't be able to assess until they happen. All algorithms are built on models of the real world. But the real world is so much stranger than we can imagine that no model yet has captured all that might impact the algorithmic thinking based on it. Models based on pure mathematics come closest-flight simulators, for example. But once one tries to model human behavior, all bets are off. In AI circles this is called the "alignment problem," the problem being that models never fully align with the realities they model.
For that reason, algorithmically controlled processes are prone to surprise us when they interact with the real world. But don't blame the algorithms-most algorithms do a good job of optimally managing the models they were programmed for. It's just that an imperfect model will cause even a perfect algorithm to screw up. The first rule of coding-garbage in, garbage out-compels it.
TerraUSD crashed because-its model of the world didn't fully contemplate both USD and BTC falling out of favor at the same time. The SEC's legitimate concern is that a similarly flawed model drives mAssets, one created, as always, by that very flawed man in the mirror.