Advertisement Opinion Matt Levine NYSE Forgot to Open Yesterday Also ESG securities fraud, SOFR adjustments and an Elon Musk capital raise that might not be oversubscribed. By Matt Levine +Follow January 25, 2023 at 11:35 AM PST NYSE The basic problem of the stock market is that a lot of people want to buy a lot of stock, and a lot of people want to sell a lot of stock, but not at the same time. I might want to buy 1,000 shares of Wells Fargo & Co. stock, and you might want to sell 1,000 shares of Wells Fargo fi ve minutes from now. Our desires match up, but not in time, so we can’t trade with each other. In theory, there are about three ways to deal with this. Method 2 is sort of the standard paradigm for how the US stock market works. It is sometimes called a “central limit order book,” or “CLOB.” It has its controversies, and lots of the market-structure stu ff we talk about around here — payment for order fl ow and dark pools and fl ash boys and blah blah blah — has to do with that process of market makers intermediating trades in time. More from Bloomberg Opinion GDP and Jobs Show the Stock Market Bears Are Still Early Abortion Pill Lawsuit Won’t Get a Fair Shake in Post-Roe America It's Not JPMorgan’s Fault If Frank Lied Boomers, Talk to Your Kids About Their Inheritance I won’t discuss those controversies here, but I do want to mention one complaint that people sometimes have, which is that liquidity in the stock market is somehow illusory or fl eeting or “not real.” Part of what this means is that, if the national best bid and o ff er for Wells Fargo is $44.99/$45.01, you could go to a market maker and buy Wells Fargo for $45.01, or sell it for $44.99, but you couldn’t buy or sell very much of it. The way the CLOB works is that market makers post limit orders at the exchange; each market maker will say “I’ll buy 100 shares of Wells Fargo at $44.99, and 200 more at $44.98, and 300 more at $44.95,” etc. And then when you come in with a market order to sell, the exchange looks at the limit orders it has on the book, and goes through them in price order to execute your trade. So if you want to sell 100 shares, you trade at $44.99. If you want to sell 1,000 shares, you trade some at $44.99 and maybe some at lower prices. If you want to sell 100,000 shares, you might run through all of the orders that the market makers have put on the order book. You might sell some at, like, $44, and then descend into weird territory. Maybe some hedge fund put in an order to buy at $20 and forgot about it, and you end up selling to them at $20. The point is that the limit order book does not represent the true economic supply and demand for a stock. It just represents the supply and demand for the stock right now , mainly from risk-averse high-frequency electronic market makers. If you want to sell 100,000 shares, you break that up into smaller orders so as not to scare o ff the market makers, you do it over some period of time, and eventually enough people will want to buy that you’ll be able to sell at a reasonable price. There will be some price impact of your trading — you can’t sell 100,000 shares at $44.99; supply and demand matter — but you won’t sell at $20, either. But every so often a trader at a big institutional investing fi rm will put in an order to sell 100,000 shares, and instead of hitting the “break this into small orders and sell over 8 hours” button she will hit the “put in a market order to sell all of this immediately” button, and the stock will brie fl y crash as her trade eats through the entire order book and ends up printing at ridiculous prices. And then the stock will recover, because its value hasn’t really changed; it’s just that there were not enough orders resting on the book to execute that trade sensibly. This is sometimes called a “fat fi nger” error , because the only excuse for it is that your fi nger is too big to hit the right button. As I said, Method 2 is the main procedure, but Method 3 is also really important. In fact there are some obvious times when lots of people all want to trade at the same time: The US stock market’s main opening hours are 9:30 a.m. to 4 p.m., and people naturally gravitate toward trading at the open or the close, at 9:30 or 4. And so the big listing exchanges, the New York Stock Exchange and Nasdaq , run opening and closing auctions for the stocks they list. The rough idea of the auction is: The opening auction is particularly interesting. Intuitively, the stock market goes to bed at 4 p.m. and wakes up at 9:30 a.m. the next day. Stu ff happens overnight: News breaks, companies announce earnings and mergers, fund managers and retail traders stay up all night reading research reports and decide to buy some stock in the morning. Most days, most stocks will open at a price that is pretty close to the price they closed at the day before. But sometimes there will be big jumps: If a company’s stock closes at 4 p.m. at $45 (i.e., the price in the closing auction is $45), and then it announces good or bad earnings at 4:15, it might open at $50 or $40 the next morning (i.e., the price in the opening auction will be a lot higher or lower). The opening auction is where the day’s price discovery happens: All the “real” investors, pension managers and hedge funds and retail traders on Robinhood, have updated their views on value overnight, and then they meet at 9:30 in the opening auction and work out what the market-clearing price is. Notice that in Method 3 you don’t really need market makers: All the “real” buyers who want to own the stock show up to buy, and all the “real” sellers who own stock and want to get rid of it show up to sell, and they all trade at the same time at a price that balances their supply and demand. Market makers are necessary in a central limit order book, because they intermediate between real buyers and sellers who want to trade at di ff erent times, but the auction solves that problem in a di ff erent way. Also, market makers probably don’t want to trade in the opening auction. A market maker does not have a deeply informed view about the fundamental value of a stock; it just tries to turn over inventory quickly and do trades at pretty close to the previous trade. At 9:30 a.m., there is no previous trade The market is just waking up, there is news to digest from the previous 17.5 hours without trading, and lots of informed fundamental investors have been digesting it and are now trying to fi gure out the right price. The market makers have no advantage, and will want to stay out of the way. And so there’s an opening auction among mostly real investors, and it sets the opening price, and a bunch of shares trade at that price, and then a millisecond later normal trading starts. And when normal trading starts, the market makers put in their orders to buy and sell stock — probably to buy it for about a penny less than the opening price, and to sell it for about a penny more than the opening price — and regular orders arrive to trade with them, and the price moves around with supply and demand, and the market operates with a central limit order book for the next 6.5 hours. Yesterday the New York Stock Exchange made one teeny little mistake: It forgot to do some opening auctions A technical glitch at the New York Stock Exchange on Tuesday brie fl y caused wild price swings and a temporary trading freeze in stocks of major companies such as Exxon Mobil Corp., McDonald’s Corp. and Walmart Inc. Hundreds of stocks experienced erroneous prices as a result of the incident, the biggest snafu to hit a U.S. stock exchange in several years, according to a spreadsheet of a ff ected securities released by the NYSE. The NYSE said in an accompanying notice to traders that, due to a “system issue,” it had failed to carry out an opening auction in a number of stocks. The opening auction is the critical process at 9:30 a.m. ET each day that determines the o ffi cial, beginning-of- day prices for stocks listed on the exchange. Without a reliable o ffi cial opening price, some stocks traded at unusually low or high prices as the trading day began. That in turn caused sharp price swings, triggering trading halts under market rules designed to curb extreme volatility. “It’s a real mess,” said Dennis Dick, founder of Triple D Trading, a proprietary trading fi rm. “I’ve traded for 22 years, and I’ve never seen the opening cross at NYSE go haywire.” Intuitively, what seems to have happened is that if you woke up at 7 a.m. and put in an order with your broker to buy 500 shares of Walmart at the open, and then you went o ff to work at your job and forgot about it, your broker probably put in a market order to buy 500 shares in the opening auction. And then there was no opening auction. So your order got sent to the central limit order book to execute as a normal trade, instead of executing in the opening auction. There are two big problems with that. One is that, at 9:30:00, the central limit order book is empty . The market makers who intermediate trades in time are waiting on the opening auction to populate the order book; if the auction doesn’t happen and CLOB trading opens anyway, there’s no one there to buy from the sellers and sell to the buyers. I mean, not literally no one — somebody has put in limit orders to buy and sell — but the professional electronic market makers whose job it is to post prices in a tight range have not really shown up yet in meaningful size. So the order book is thinly populated, and any new orders — like your order to buy 500 shares — can move the price more. The other problem is that the orders in the opening auction can be big . If you are looking to buy 10,000 shares at the market price at 11:15 a.m., you will break that into smaller orders and be tactical about how you trade it, because you know that sending a 10,000-share market order can crash through the order book and get you a terrible price. But if you are looking to buy 10,000 shares at the market price at 9:30 a.m., you might just put in a 10,000-share market-on-open order, because you know that lots of people want to trade in the opening auction and so you will get a good price. If there is no auction, and your 10,000-share market order gets sent to the regular order book instead, then NYSE has e ff ectively fat- fi ngered you. It sent your unusually large order into an unusually thin order book. And so the stock crashes, or soars, depending on whether you are buying or selling. The details are still a little vague and I am being approximate here. You might ask: “Well, if there are a lot of orders to buy a lot of shares and a lot of orders to sell a lot of shares and they all got sent to the regular order book at the same time, wouldn’t they trade with each other and cancel out and lead to pretty normal prices,” but that is apparently not quite how it works. The thing about an auction is that all the orders execute at once; in a central limit order book they execute in some sequential order. You could imagine NYSE looking at its list of orders, seeing a market sell order, executing it against the few resting limit buy orders, crashing the price down, looking at the next order, seeing a market buy order, and executing it against the few resting limit sell orders, shooting the price right back up again. And that’s how you get “sharp price swings, triggering trading halts.” Anyway in the grand scheme of things this doesn’t matter: Yesterday’s opening did match buyers and sellers, over time, and fi nd the market price, just in a harrowing and chaotic way. If you traded at the wrong price, it mattered to you, since you made or lost money, though a lot of those trades will be busted because they were “clearly erroneous.” NYSE has a sensible market structure for its 9:30:00 a.m. trading, and a sensible market structure for its 9:30:01 a.m. trading, and they are very di ff erent , and if you mix them up you get a mess. Everything is securities fraud Around here I like to say that every bad thing that a public company does is also securities fraud. It does the bad thing, it does not immediately tell shareholders about the bad thing, later the shareholders fi nd out about the bad thing, the stock drops, and the shareholders sue, saying “we were tricked into buying your stock because you lied to us about not doing the bad thing.” This is I think a broadly correct description of how US securities class actions work in practice, but it is not a technically accurate description of the law, and it is missing some nuance. For one thing, simply not mentioning a bad thing might not be enough to create securities-fraud liability: To sue and win, shareholders will need to point to some misleading statement that the company did make. (So sexual harassment might be securities fraud if a company has a stated policy forbidding sexual harassment, or a risk factor in its annual report saying “we rely on the services of our chief executive o ffi cer and would have problems if he was a sexual harasser” but not mentioning that he is.) For another thing, there will be arguments about whether the misstatements or omissions were material to investors. If shareholders say “we bought your stock because you lied to us and told us you were ethical, and we believed you,” a plausible answer would be: “No you didn’t, you didn’t care about our ethics, they don’t matter to our stock price and you were not actually defrauded.” (This was roughly the issue in a 2021 Supreme Court case about Goldman Sachs Group Inc. and the “everything is securities fraud” theory.) One way to synthesize these points is that there are some areas of corporate behavior where a company can be good or bad , but these areas are not relevant to shareholders . That is: Like one bad thing that a public company could do is that its executives could have really bad fashion sense and dress really badly. (Not a high fashion company I mean, just like a software company or whatever.) That would be bad , in some aesthetic sense, but it would not be securities fraud . Investors would not make investing decisions based on executive fashion, the company would not make any claims about it in its fi lings, and if a photo emerged of a badly dressed executive the stock would not drop. These categories are fuzzy, though; there are some categories that investors once did not care about but now do, some sorts of behavior that sometimes cause stock drops and sometimes don’t, some things that are obliquely hinted at in securities fi lings. Sexual harassment scandals, for instance, seem more likely to cause business consequences and stock drops now than they did 10 years ago, fi nancial decision-makers ask about them more frequently , and they are probably covered by companies’ disclosed ethics policies. You could argue that 10 years ago executive sexual harassment was bad but not securities fraud, but now it is also securities fraud. One rough way to think about the “everything is securities fraud” theory is that, once upon a time, most sorts of corporate behavior fell into these categories — investors were presumed not to care about them, they weren’t discussed in fi lings, etc. — and the only area that mattered was, like, fi nancial results. “Securities fraud” meant lying about earnings. And then over time various areas of corporate behavior — executive ethics, cybersecurity, etc. — became things that companies disclosed and investors were presumed to care about, and so now shareholders can sue companies if bad things happen in those areas. Surely the biggest and most important example of this shift is ESG. Stereotyping very crudely, once upon a time investors did not care about the environmental, social and governance behavior of public companies, and now they care very much. For many investors, these are the most important factors; many investors have “ESG” (or “sustainable,” etc.) right in their fund names. And companies produce increasing amounts of disclosure about ESG, in part because that’s what investors want but in large part because it’s what regulators demand And so you could imagine 20 years ago thinking “companies produce fi nancial statements, investors read the fi nancial statements and use them to make investing decisions, and if the fi nancial statements are wrong that is securities fraud.” And now you could imagine thinking “companies produce ESG statements, investors read the ESG statements and use them to make investing decisions, and if the ESG statements are wrong that is securities fraud.” The ESG statements haven’t replaced the fi nancial statements, but they exist parallel to them, and they are ... perhaps not equally important, but certainly Advertisement I could just wait. I could put in an order on the stock exchange to buy 1,000 shares, and then it could wait around for fi ve minutes until you put in an order to sell 1,000 shares, and then we could trade with each other. Sometimes it might take days. Lots of markets work kind of like this, and there are aspects of this in the US stock market, but this is mostly not how US public stock markets work. When I put in an order to buy Wells Fargo stock, it gets executed more or less immediately. 1 1 There could be market makers . Some people — banks, high-frequency electronic trading fi rms — could be willing to buy from all the sellers and sell to all the buyers, intermediating trades in time. I put in my order to buy 1,000 shares, and some electronic market maker instantly sells to me at $45.01 per share; fi ve minutes later, you put in your order to sell 1,000 shares, and the electronic market maker instantly buys from you at $44.99. We have each e ff ectively paid a penny per share for “immediacy,” the service that the market maker provided of letting us trade instantly instead of waiting to fi nd each other. The market maker is in some sense not a real investor like you and me, and it will not generally have some deeply informed view on the value of the stock: Its job is to buy and sell quickly, turn over its inventory, and get paid the spread between the $45.01 it charges me and the $44.99 it pays you. 2 2 The exchange could get us all together at the same time. The stock exchange could say: “Hey, everybody who wants to buy or sell Wells Fargo shares, we’re gonna do a big share swap at 9:30 a.m. tomorrow. Send in your orders, and we’ll match them up and let everyone trade at the same price at the same time.” 3 3 Advertisement 4 Everyone who wants to buy or sell stock at the open or the close puts in orders to buy or sell, in the minutes leading up to 9:30 or 4. Some of those orders are market orders (“I want to buy/sell stock at whatever the auction price ends up being”), while others are limit orders (“I want to buy stock in the auction if the price is $44.95 or less”). The exchange’s computers fi nd a price that matches up the most orders and chooses that as the auction price. Everyone who put in an order to buy at the auction price or more (or with a market order) buys, and everyone who put in an order to sell at the auction price or less (or with a market order) sells. 5 Advertisement 6 Advertisement The shareholders do not care about these areas of behavior; they do not evaluate whether a company behaves well or badly in these areas, and those questions do not inform their buying and selling decisions. 1 Companies do not talk about these areas of behavior in their public fi lings. 2 If it turns out that a company is bad in these areas, the stock doesn’t usually drop. 3 Advertisement Why nuclear is the next frontier in green finance. Ideas that make a difference. Read now Industries in transition Read More Financing critical minerals on the road to a more sustainable auto sector READ MORE Share Subscribe