| Programming note: Money Stuff will be off again tomorrow, back on Monday. Speed bumps The paradox at the heart of market structure is that "Standing Limit Orders Are Trading Options That Offer Liquidity."[1] If a stock is trading at $100, and I place an order to buy it at $99.99, that order goes onto the order book, and I am bound by it. I have effectively written a put option on the stock. If the value of the stock goes down to $99.95 before I cancel my order, the put will be exercised against me: Someone will hit my bid, and I will be forced to buy the stock at $99.99. If the value of the stock goes up to $100.05 the put will not be exercised, and I will not get any stock. My risk is asymmetric: I buy the stock when it's going down, but I don't buy it when it's going up, just like with a regular put option.[2] If I am a market maker, standing ready to buy or sell the stock, it's even worse: When the stock is at $100, not only do I put in a bid to buy it at $99.99, I also put in an offer to sell it at $100.01. I've written a call option: If the stock goes up to $100.05, I will sell it at $100.01. I have risk in both directions: If the stock is going down I'm buying it; if it's going up I'm selling it. If I actually sold a put option, I would get paid for it: Someone would pay me a premium in exchange for this option to sell me stock at a fixed price. Same with a call option. Options are valuable, and modern financial math lets us calculate their value. But if I just put a buy and a sell order on the stock exchange, no one is paying me a premium for them. I am giving away free options. The options aren't worth that much: I can cancel them at any time before exercise, which means they have a theoretical expiration date of a fraction of a second, which reduces their value to nearly zero. But not quite to zero. By writing these options, I am providing value to other people—people who have the opportunity to trade with my orders and exercise the options against me—but no one is paying me for it. It seems irrational. But it is good. It is good for a stock exchange to have orders: It is useful to have a lot of people sitting around bidding to buy a lot of stock at $99.99, and a lot of other people sitting around offering to sell a lot of stock at $100.01, because then if someone wants to buy (sell) a lot of stock in a hurry they can easily and instantly do it at $100.01 ($99.99). The stock exchange, as the saying goes, provides liquidity. It lets real investors execute their trades instantly. That's what the exchange wants to do; that is the business it is in. So it wants to attract market makers, and it wants to attract a lot of standing limit orders. It wants people to provide a lot of free options on its platform. It wants to cater to those people, to attract and assuage them, to make them more willing to do this irrational-seeming thing of providing free options. And so many stories about market structure are basically stories about finding ways to compensate market makers for the otherwise irrational-seeming decision to write free options to the rest of the market. So "high-frequency trading" is in large part a story about market makers building very fast computer systems that enable them to leave their free options outstanding for as little time as possible: Unlike a real option, orders can be canceled at any time, so if you are the fastest to update your prices you can avoid buying stock at $99.99 when its value falls to $99.95. (If you know you can always cancel when the price moves, before anyone else notices the move, then the option really is worth zero.) The many complaints in Michael Lewis's book "Flash Boys" about liquidity disappearing when hedge fund managers try to buy a lot of stock are complaints about exactly this: market makers being faster to cancel their options than investors are to exercise them.[3] Similarly "payment for order flow" is mostly a story about market makers trying to offer their options only to uninformed retail investors who will not exercise them optimally, making the options cheaper. "Maker-taker pricing" is the most straightforward story of all: It is a story about stock exchanges charging investors for exercising the options, and using some of the money to pay market makers for writing the options. The options are valuable, so why shouldn't their buyers pay for them and their sellers get paid? Then there is "last look." Last look is a feature of some trading systems—it has been particularly popular in electronic platforms for foreign exchange trading—that allows market makers to cancel their orders for a split second after they match. Or, in options terms, it allows market makers to cancel the options they wrote for a split second after someone exercises the options against them. That pretty clearly reduces the value of the option to near zero. This makes it much more attractive for market makers to provide orders on those platforms. By attracting a lot of orders at relatively tight bid/ask spreads, the platforms provide a lot of value to real users (people who just want to buy and sell FX immediately). On the other hand, letting the market makers cancel the orders reduces that value: The people who want to buy and sell FX immediately will get frustrated if their trades keep getting canceled by last look. There is a balance; you want the market makers to be able to cancel when the market moves hard against them (which lets them reduce their risk), but you don't want them to cancel every time they get hit (which would be pointless for them and for the end users). Also last look just sort of looks funny and has been controversial. It seems maybe a bit too obviously like cheating, giving market makers an opportunity to walk away from their commitments. But then there are "speed bumps," which are features of an exchange that allow some order actions to move faster than others. The idea of the speed bump comes from IEX, the Investors Exchange, which is profiled in "Flash Boys." In "Flash Boys," IEX are the heroes and electronic market makers who cancel orders are the villains, and so as you'd expect IEX's speed bump was not especially designed to protect market makers. Quite the reverse: IEX's original use of the speed bump was to increase the value of options offered by market makers, by making it harder for them to cancel their orders when the market moved against them. (The speed bump effectively slowed down the market makers' cancellations of orders but did not slow down the liquidity-taking orders that real investors used to exercise options against the market makers.) This was controversial, and IEX ultimately had to give up some of the features of its speed bump to be approved as an exchange. But it kept the core of the speed bump, and that gave other exchanges ideas. The idea was basically: Instead of a "last look," which is controversial everywhere and effectively illegal in public U.S. equities markets, why not have an asymmetric speed bump that protects market makers. You delay any liquidity-taking order, but you don't delay market makers' ability to modify their existing limit orders. So if a stock is at $100 and a market maker has a bid and offer of $99.99/$100.01, and there is some news that would push the price down to $99.95—say, a trade on another exchange, or an earnings release, or whatever—then the market maker gets a split second to change its bid and offer to $99.94/$99.96 before anyone else has a chance to buy from it at $99.99. My basic theory of this sort of speed bump—not the IEX one, but the ones embraced by other exchanges—is that they are a socially acceptable substitute for last look. They're a way to let market makers provide bids and offers for a stock (or option or currency or whatever), but get away from those bids and offers if the price moves against them; a way to reduce the value of the options that market makers are providing for free, in order to encourage the market makers to provide more of them. As a substitute for last look, speed bumps have obvious advantages. They look fairer, for one thing: The market makers aren't cancelling their orders because they got hit, but only because there is some external news that makes them think the market moved against them. They have the imprimatur of IEX and "Flash Boys," for another thing: If you think high-frequency trading is sort of evil, then you might think that "speed bumps" are good, even though these speed bumps have the effect of protecting one kind of high-frequency trader (electronic market makers) from another (liquidity-taking HFTs). Also they are legal in U.S. equities markets: Securities and Exchange Commission rules about fair market access would probably prohibit last look,[4] and before IEX's approval as an exchange many people thought those rules prohibited speed bumps as well. But everyone likes IEX, so IEX's speed bump got approved, which makes it hard to block everyone else's speed bumps. So here's a story about how speed bumps are everywhere now: Exchanges on both sides of the Atlantic are increasingly embracing the mechanisms, which impose a split-second delay before executing trades. By 2020, more than a dozen markets in stocks, futures and currencies from Toronto to New York to Moscow will slow trading via speed bumps or similar features, if all of the currently planned launches occur. Five years ago, only a few markets had speed bumps. … Most of the latest speed-bump plans have a similar, "asymmetrical" design, meaning they don't apply equally to all trades. Such speed bumps are typically meant to favor market participants that publicly quote prices on an exchange, rather than those that try to buy or sell using those prices. The article frames this as "defying high-speed trading giants that account for a huge portion of their volume, in a bid to appeal to more-traditional clients," but I am not sure that is the whole story. Pension funds are not naturally in the business of publicly quoting prices on exchanges and updating them every millisecond! I think a more accurate framing might be that exchanges are defying one class of high-frequency trader in a bid to appeal to another class of high-frequency trader, the kind that posts lots of bids and offers but wants to avoid adverse selection. That kind of high-frequency trader—the kind that provides liquidity, even if that liquidity might sometimes be illusory—is the kind that exchanges really like. Is this good? I dunno. The paradox is not just that market makers provide valuable options for free; it is also that exchanges fight to get more of those valuable options by making them less valuable. If you reduce the value of the options to zero, but get a lot of them, have you accomplished anything? If you get lots of standing limit orders on your exchange, but market makers cancel those orders every time the market moves, are you really giving real investors any valuable liquidity? My sense is that the answer might be yes, that the exchanges have a plan here, that there is a value in attracting a lot of orders even if they are cancelable, that liquidity begets liquidity, even if some of the liquidity is illusory. But I'm not sure the answer is obvious, and real investors might rather have a world where market makers bid $99.97 and offer $100.03 but are held to those prices than a world where they bid $99.99 and offer $100.01 but can walk away from them. That is not the direction we're moving in though. Overstock We talked on Tuesday about Overstock.com Inc.'s airdrop of blockchain stock onto its shareholders. For every 10 shares of regular, Nasdaq-traded, normal Overstock common stock, Overstock will distribute one share of "Digital Voting Series A-1 Preferred Stock," a class of stock that has the same voting and economic rights as the common stock but that can be traded only on Overstock's weird blockchain securities platform. I thought this was kind of cool; I am all for experimentation, and this is an interesting way to force a lot of shareholders to start messing with blockchain whether they want to or not. But I missed something about the Overstock airdrop, something that Paulo Santos caught and wrote about at Seeking Alpha, something that I am frankly embarrassed not to have figured out myself. Patrick Byrne, the chief executive officer of Overstock, has been on a years-long war against short sellers. He is a true believer in the theory that "naked short sellers"—people who illegally sell his stock without owning or borrowing it—have conspired to drive down his stock price, and part of his fondness for blockchain stems from his (probably accurate) belief that a blockchain-based securities market would make naked short selling impossible. "Our legacy OSTK shares trade in a capital market with trading and settlement mechanisms about which I have long made my criticisms and doubts known to the public," said Byrne on Tuesday, "whereas our new blockchain-based A-1 shares trade in a blockchain-based capital market which I believe is resistant to such dynamics." I suggested on Tuesday that the dividend of blockchain stock might be the first step in Byrne's plan to, ultimately, move all of Overstock's stock trading off the regular capital markets (where he is beset by naked short sellers) and onto his special blockchain (where he wouldn't be). That still seems correct to me. But what I missed is that the blockchain-stock dividend punishes actual short sellers of Overstock's regular stock right now. Here's Santos: As with dividends, any other asset spun off from the original shares has to be provided by the short seller to the holders of the shares he sold short. It's here that Overstock's creativity came in. What did Overstock just do? Well, it decided to create and award a dividend in something the short sellers don't have for delivery: A digital share dividend. The "Digital Voting Series A-1 Preferred Stock (the "Series A-1")." These shares are supposedly traded in on the "PRO Securities alternative trading system" operated by Overstock's tZERO subsidiary. Thing is, of course, no normal broker actually trades on this "market" as of today. Basically, the issuance of this digital asset now forces Overstock short sellers to turn around, register on this alternative market through deeply unknown brokers, and try to get some of these tokens to cover what they need to deliver. If you are a short seller of Overstock stock, you have borrowed that stock from someone, and your borrowing arrangement obligates you to pay back any dividends or distributions paid on the stock. If Overstock pays a dividend of $1, you have to pay your share lender $1. If Overstock pays a dividend of one share of Digital Voting Series A-1 Preferred Stock, you have to give your lender one share of Digital Voting Series A-1 Preferred Stock, which means you have to go and buy it. This presents problems for you. For one thing, Overstock's blockchain market is maybe, as of today, slightly less robust than Nasdaq? Less popular among traditional brokers, less liquid, harder to use, that sort of thing. For another thing, if you are short Overstock it is presumably because you don't … like … Overstock. You have doubts about the business, but also specifically about Byrne, and about his pivot to blockchain. So figuring out how to buy and deliver Overstock stock on its own blockchain is going to be particularly off-putting to you. You're going to have to sign up as a customer for Byrne's weird project, the one you are shorting. Worst of all, "these new shares of Series A-1 to be issued in connection with the Dividend have not been, and are not required to be, registered under the Securities Act of 1933," says Overstock, and so they can't be sold for six months. If you are a short seller who has borrowed a bunch of Overstock stock, now you have to go out and buy Overstock blockchain stock, but the people who are getting it are not allowed to sell it to you. Seems challenging! If I were an Overstock short seller, I'd be mad, but I'm not, so I'm impressed. This is a good move! Like, Byrne definitely wins this round against the shorts, you know? It feels a little like cheating insofar as it only punishes legitimate short sellers: If you borrowed Overstock stock to short it, now you have to go find Series A-1 stock to deliver to your lender, but if you were actually naked short Overstock stock, then this is no problem for you since you have not borrowed stock and have no obligation to deliver anything.[5] But as a pure tactic to deter and punish short sellers it is rather nifty. It makes short sellers' lives hard, not by doing anything to increase the long-term value of the stock and thus make their thesis wrong, but purely by adding technical difficulties to maintaining the short. It sends a message to short sellers: This trade will be painful for you even if you're right, so stay away. If I were Byrne I would be pitching Elon Musk hard on listing Tesla Inc. on the Overstock blockchain trading platform and pulling this sort of move. I bet a lot of Tesla long investors would be happy to follow Musk onto the blockchain—"Crypto Traders Also Enjoy Investing In Tesla, Claims Redundant Study," was the Dealbreaker headline a while back—and it would make his short sellers miserable. It's a natural fit! Carlyle It used to be that the big stock indexes all included companies with dual-class stock. Alphabet, Facebook, Berkshire Hathaway, these are all giant famous companies that have one class of shares with more votes than another class of shares, and they are all in the S&P 500 Index and many other indexes. But in recent years, inspired by Alphabet and Facebook, more and more companies—particularly big tech startups with charismatic visionary founders—went public with dual-class stock that allows the founders to keep control of their companies indefinitely. Investors grumbled ineffectually about this until Snap Inc. did it in an absolutely ridiculous way, selling public shareholders common stock with zero votes while keeping all of the voting power for founders and early investors. Investors went and bought Snap's stock anyway, but they grumbled about it more effectually, and the result was that several of the big index providers announced that they wouldn't accept any new dual-class stocks in several of their big indexes. In particular, S&P Dow Jones Indices won't add dual-class stocks to the S&P 500 anymore, though Alphabet and Facebook and Berkshire can stay. I thought that was big news. Before Snap, when a company was thinking about an initial public offering, its founder would say to the bankers "I would like to control my company indefinitely using dual-class stock like Mark Zuckerberg did," and the bankers would say "well investors do not like that so there will be some discount to the price if you insist on it," and the founders would say "really?" and the bankers would say "no, not really, the investors will complain but they will put in absolutely the same orders at the same price with or without dual-class stock." I mean, I don't know, the bankers probably didn't say it like that, but it was obviously true, and the founders would notice that. So the bankers' warnings and the investors' grumbling didn't have much effect. But after Snap, bankers can say "look if you have dual-class stock the S&P 500 index funds will never buy your stock, and index funds are a huge and growing percentage of the market, and yes sure right you won't be in the S&P 500 on day one anyway, but the fact that you'll never be in the index means that your long-term value will be lower, which means that rational investors will not be willing to pay as much for your stock in the IPO as they would with single-class stock." I assumed that would sort of be the end of big dual-class IPOs. But, nope! Everything is still dual-class. Lyft. Pinterest. Spotify (not an IPO but still). Chewy. Chewy! Not a charismatic-visionary-founder company, but an IPO of a private-equity-backed company; it just went with dual-class because everybody does. On the other hand: Carlyle Group LP said Wednesday it would abandon its partnership structure and become a corporation with a single class of shares, going a step further than private-equity peers that have already converted. Each of the new Carlyle shares will have one vote, giving greater say to the roughly 30% of shareholders who aren't insiders at the firm. The changes are expected to pave the way for Carlyle's inclusion in indexes such as FTSE Russell's, which have minimum requirements for public-shareholder voting rights, and the S&P 500, which doesn't allow companies with more than one class of shares. Carlyle is a pretty natural case for a dual-class stock. It is not exactly run by its charismatic founders, but they're still around. It has fiduciary duties to the investors in its private equity funds, duties that might conflict with the desires of shareholders of the management company, which provides a good reason for the managers of that company to want to freeze out shareholders. On the other hand it's a public company with an $8 billion market capitalization and $46.6 billion of assets under management. At some point—at some fairly mature point, when its founders have stepped back from day-to-day management, like now—maximizing the market value of the company by making it index-eligible becomes more important than maintaining control. (Also, to be fair, the insiders still have a majority.) And so Carlyle will flip the switch. Maybe that is actually the effect of the indexes' ban on dual-class stocks. It won't affect brash young founders at the time of the IPO: The dual-class discount is hazy and unknowable for a stock that isn't public yet, and a new IPO company won't join the S&P immediately anyway so it's not like you're explicitly losing out on index-fund orders. But critics of dual-class stock tend to focus on criticizing permanent dual-class stock, and on proposing that new public companies have sunset provisions getting rid of the dual-class stock after some time. Perhaps that's not necessary; perhaps the market—well, the index providers—will take care of it. When they're young and hungry and active, the founders will value control most; as they and their companies age, they will value valuation more, and at some point the lines will cross and they'll voluntarily flip into single-class structures. Utility token When we first met George Weiksner, early last year, he was 11 years old and hanging out at "the upscale New York City restaurant Bagatelle, which hosts Crypto Mondays, a meet-up group where crypto and blockchain enthusiasts meet, glad-hand and share ideas about the world of digital currencies." He had his own "universal cryptocurrency for games," of which he was the chief executive officer, and it had the delightful name Pocketful of Quarters Inc. I wrote at the time: I don't know about his cryptocurrency but I will say it's a pretty good name! If you got a 26-year-old to name it he'd probably have come up with, like, "GameCoin." "Pocketful Of Quarters" is unexpected but right, evocative and memorable. More cryptocurrencies should be named by 11-year-olds is I guess the lesson here. Well, they should. On Friday the Securities and Exchange Commission released a no-action letter allowing Pocketful of Quarters to do an offering of its tokens—which it, naturally, calls "Quarters"—without registering them as securities. We have talked a few times about the SEC's criteria for deciding whether or not a crypto token is a security, and it's super clear that Quarters aren't. They're not a fundraising mechanism, for one thing: As the SEC points out, "PoQ will not use any funds from Quarters sales to build the Quarters Platform, which has been fully developed and will be fully functional and operational immediately upon its launch and before any of the Quarters are sold." Nor are they a speculative investment: "Quarters will be made continuously available to gamers in unlimited quantities at a fixed price." Instead they are a pure utility token: "The Quarters will be immediately usable for their intended purpose (gaming) at the time they are sold," and "PoQ will market and sell Quarters to gamers solely for consumptive use as a means of accessing and interacting with Participating Games." Only game developers—and, ominously, "Influencers"—will be able to redeem Quarters for cash (or rather Ethereum), "at pre-determined exchange rates by transferring their Quarters to the Quarters Smart Contract." Quarters are meant to be a way for game developers on the Pocketful of Quarters platform to build games and accept payment from gamers. They could just use dollars, I guess, but, shh, shh, smart contracts via crypto make this super doable. Honestly all of this is pretty obvious; as SEC Commissioner Hester Peirce has pointed out, if the SEC declared that this sort of thing was a security, then it would have to argue that actual arcade tokens were securities, and that way madness lies. But I do want to say that it's particularly obvious in this case because the kid gave his crypto a really good name. If he'd called the thing "GameCoin" then there'd be a long and tedious argument about whether its investment or utility aspects dominated, and his lawyers would argue that they were "like arcade tokens that can be used on a variety of unrelated games but cannot be used outside the arcade" (as they actually did) while skeptical critics would reject the analogy. But if you call it "Pocketful of Quarters" then everyone is immediately going to be like "oh right they're arcade tokens." Similarly, a very good argument against classifying them as securities is that no one will buy them for price appreciation. Having a mechanism to sell and redeem unlimited quantities at a fixed price is of course useful support for this argument, but so is calling them "Quarters." No one is going to buy a "Quarter" hoping that it will be worth $1,000; the fixed price is right in the name. Anyway my point here is that all of crypto should be run entirely by tweens. Things happen Texas Pacific Land Trust Enters into Settlement Agreement with Investor Group. (Earlier.) Sports Cars, Psychopaths, and Testosterone: Inside the New Frontier of Fund Manager Research. Negative-Yielding Debt Pile Hits Record $14 Trillion as Fed Cuts. London Stock Exchange clinches acquisition of Refinitiv for $27bn. Barclays outperforms US rivals with jump in bond trading revenue. FBI Examining Possible Data Breaches Related to Capital One. Jeffrey Epstein Had a Door Into Apollo: His Deep Ties With Leon Black. The College Financial-Aid Guardianship Loophole and the Woman Who Thought It Up. People are worried about stock buybacks. CFTC Charges Trader with Spoofing in Financial Futures Markets. Why Martin Shkreli Lost His Appeal. What René Girard Can Teach Us About Bubbles. How To Rob a Bank. I'm Sorry to Report There's a Bitcoin CBD Gym Now. Here Comes Big Boy!!!!!!! If you'd like to get Money Stuff in handy email form, right in your inbox, please subscribe at this link. Or you can subscribe to Money Stuff and other great Bloomberg newsletters here. Thanks! [1] This is the header of Section 4.4.2 of Larry Harris's textbook "Trading & Exchanges: Market Microstructure for Practitioners," which is where I learned everything I know about market structure. [2] This is sometimes called "adverse selection." [3] Other stories about high-frequency trading, on the other hand, are stories about other traders building very fast computer systems to exercise options against market makers before the market makers can cancel them. It's a race. [4] The rule requires exchanges to "immediately and automatically" respond to incoming orders without human intervention; pinging the party that entered the order to see if it still wants to go through with it seems bad. [5] Not legal advice. Obviously if you have been naked short for months you are doing something wrong and you have some obligations to somebody. In general I am a skeptic of naked-short conspiracy theories and don't think that this is actually a thing, but clearly Byrne disagrees. | 
Post a Comment