Prediction market liquidity | There are two basic ways to use computers to trade stocks: - There are what you might call “naive market making” and “simple arbitrage.” Sometimes a stock index future trades at $100 and an exchange-traded fund on that index trades at $100.01, so you buy the future and sell the ETF and collect $0.01. Or some stock last traded at $30, so you post orders to buy it for $29.99 or sell it for $30.01, hoping to buy from sellers and sell to buyers and collect a small spread. You program computers to do this quickly, make small amounts of money on each trade and try not to lose money. You and your computer do not have the time or inclination to think about fundamental value: You don’t know if the stock should trade at $20 or $30 or $40; you just know (or can confidently predict) that you can buy it at one price and sell it at another price relatively quickly.
- There is what you might call “quantitative investing,” where you program a computer to look at some set of data about a stock and figure out what it’s worth. If it last traded at $30, and your computer tells you it’s worth $40 or $31 or perhaps even $30.05, you (or your computer) buy the stock and hope to sell it later at a higher price when the market comes around to your computer’s view.
These things bleed together; a human market maker at a big bank, or an electronic market-making algorithm at a big electronic trading firm, will probably have some notion of fundamental value. It is possible to do market making or arbitrage with no concept of fundamental value, but in a competitive market it is hard and risky. If you are buying a stock at $29.99 and selling at $30.01, and some well-informed investor knows that it’s really worth $40, she will keep buying from you at your price and you will lose a lot of money. (This is called “adverse selection.”) The only safe ways to do market making, really, are: - Have an informed and sophisticated view of the “correct” price, so you do not constantly face adverse selection, or
- Trade against idiots.
A lot of very smart people devote their lives to the first approach, but it is important to point out that a lot of very smart people also devote their lives to the second approach. Big electronic proprietary trading firms understand that the public stock markets are full of sharks, that is, sophisticated investors with informed views on prices who will pounce on them if they get their prices wrong. One thing that they do about this is try to make their prices right, but another thing that they do is try to avoid the sharks. We talk sometimes around here about “payment for order flow,” which is the mechanism that US stock market makers use to avoid sharks: A market maker will pay a retail brokerage to trade with its customers’ orders, because those customers are unlikely to be especially well informed about the correct stock price. Market makers will pay more to brokers whose customers are particularly uninformed. Uninformed orders are what you want. Taking a step back, though: There are idiots everywhere. [1] If you are fundamentally in the business of finding ways to trade with uninformed people, there is no sense limiting yourself to the stock market. One reason that several of the big market-making firms jumped enthusiastically into crypto — or their employees left to jump into crypto — is, uh, you know. Not so much “crypto is full of idiots” (maybe!) but more “a lot of cryptocurrencies trade frequently at large market caps with no particular fundamental value, so you are unlikely to be picked off by sharks with a more sophisticated view of fundamental value than you have.” The hot financial thing in 2026, though, is prediction markets. Prediction markets seem to have a lot of the same appeal to trading firms as crypto markets: - They attract a lot of hobbyists and gamblers rather than high-powered professional investors, so chances are good that you will not be adversely selected by sharks.
- They are small and new enough that the arbitrages have not all been competed away; it is at least possible that you’ll find the same contract trading at 55% on Kalshi and 50% on Polymarket and you can make a quick profit.
- They are global and often pretty lightly regulated, and philosophically (and technologically) often pretty crypto-y, so you can apply your crypto learnings to trading predictions.
- They are plausibly uncorrelated to your stock trading business.
- Your interns will love it. Just, you know: One reason you got into this business was to make steady profits, but another reason you got into it was to have a little fun, and arguably “we build models to gamble on sports and awards shows” is more entertaining than “we build models to arbitrage across US equity trading venues.” It’s the Wild West of modern markets, plus it’s something you can talk about at parties with non-financial people and maybe they’ll be interested.
Prediction markets do, however, have one big difference from crypto: There is absolutely a fundamental value, and you will find out in the fairly short term if you are right or wrong about it. The contract for “Amy Poehler will win the Golden Globe for Best Podcast” would pay out $1 if she won and $0 if she didn’t, and it resolved on Sunday. It’s not like Dogecoin, where you can perpetually and unquantifiably make the case that it’s worth whatever you want it to be. The prediction either comes true or it doesn’t, in the near term, and the contracts pay out accordingly. This does not necessarily matter directly to market makers, because they can trade in the even shorter term: If you buy that contract at $0.60 and sell it at $0.62 a minute later, you make a profit, whether or not she wins. If you are trading against pure degenerate gamblers with no insight into who will win, your flow is random and you’ll make a spread. But it’s a market where you can get picked off. If you sell that contract at $0.62 to someone who knows it’s going to $1, and then you sell them some more at $0.64 and $0.66 and $0.68, you can get in trouble. So you could imagine a path in which: - Big Wall Street proprietary trading firms get into prediction markets, because they are fun and lucrative and they can apply their skills in new territory.
- By “apply their skills,” I mean mostly arbitrage and market making against relatively unsophisticated retail hobbyists.
- But, because prediction market prices depend on fundamentals, they will occasionally get picked off by people who know more. Those people might be insider traders, or they might just be smart hard-working people who seek out fundamental information and/or build good models to predict events.
- And, as big proprietary trading firms do more prediction-market trading, that will happen more often. One reason that people do not regularly make millions of dollars insider trading in prediction markets is that nobody will sell them millions of dollars’ worth of contracts: Prediction market liquidity is mostly pretty thin. But as big prop firms get into these markets, they will provide more liquidity, which means that it will be more lucrative to pick them off.
- Which means that, over time, the big prop firms will end up investing, not just in quick arbitrages of prediction markets, but in building fundamental models. They will hire astrophysicists to program computers to use deep learning models to predict which podcasts will win Golden Globes, etc., so that they can make more informed markets in who will win the Golden Globes.
- This will create more liquidity, and more incentives to trade against it, and so big hedge funds will also build their own more medium-frequency quant predicting businesses.
One possible end state here is “the future will become transparent to us, because Jump Trading and Susquehanna and Jane Street and Citadel and Millennium and Point72 will build sophisticated models that accurately predict the outcomes of all interesting events to make their prices correct.” That is, market signals will lead very well-capitalized and highly motivated players to build perfect oracles. This is not the most likely outcome — I mean, it hasn’t really happened yet for stocks — but maybe. Another possibility is that if you are a really close follower of celebrity gossip and awards-show buzz, you might have a lucrative future on Wall Street. I wrote once about meteorologists getting hired at hedge funds: The nicest thing you can say about the financial industry is that it lavishly rewards correct understanding of the world. If you know a thing, and other people do not know it, financial markets provide an efficient and fairly general way to turn your knowledge into large amounts of money.
That was broadly true at the time, but in limited domains: Weather is relevant to all sorts of commodity prices, but podcast awards buzz mostly was not. Now it is! Anyway here’s a Financial Times story about the early stages of this: Trading groups are expanding into the rapidly evolving realm of prediction markets, hiring traders to arbitrage fleeting price discrepancies between contracts for events such as football games and elections. ... Don Wilson’s DRW is looking for a trader who will be paid a base salary of up to $200,000 to “monitor and trade active markets in real time” across Polymarket and Kalshi, according to a job advert posted last week, as it builds a “dedicated prediction markets desk”. Options trading giant Susquehanna is on the hunt for traders to “detect incorrect fair values” and identify “unusual behaviours” and “inefficiencies” on prediction markets, as well as people to work on its dedicated sports trading desk. Crypto hedge fund Tyr Capital hopes to hire a prediction markets trader “who is already running sophisticated strategies”. … Analysts said strict risk controls meant trading businesses would probably avoid placing direct bets on questions such as when US President Donald Trump will buy Greenland or which film will win the most Oscar awards in March. A more attractive proposition is arbitraging between markets offering different prices for similar outcomes — mimicking how high-frequency traders exploit spreads on different stock exchanges. “The big guys are going to be trading one market versus another, they’re not going to be throwing darts at a dartboard, betting that Trump will invade whichever country,” said Joseph Saluzzi, co-founder of Themis Trading. “In a market like this that’s so new, where different platforms are so siloed, there will be so many arbitrage opportunities,” he added. Yes, that is the sensible first step. But eventually, if you can make a lot of money on Trump invading some country, somebody will. Incidentally I have gotten through this whole column without saying “obviously this is like 90% sports gambling,” but obviously this is like 90% sports gambling. People are worried about stock buybacks | My basic model is that: - Companies try to maximize profits, and then
- They take those profits and try to use them in the highest-value way.
That is, it is a sort of two-step model: The company has an operating business, and the goal of the operating business is to bring in as much profit as possible. And then the company has a capital-allocation function, and if it makes a profit, it will allocate that money to projects with the highest expected return. Sometimes those projects might be building a new factory, or expanding the sales force, or otherwise reinvesting in the current operating business. Other times, they might include getting into a new line of business or acquiring other companies. And then there is the default option, which is: “We have a good business, it makes money, we’re happy, but we’re doing it as fast as we can and it happens to make more money than we need in the business. So we’re going to give the money back to the owners of the business, and they can figure out how to reinvest it, or use it to pay their mortgage for that matter.” Historically that was the normal thing that businesses did with their profits: People invested in companies because they expected the companies to make steady profits and give the profits to investors. (This was called a “dividend.”) Modern stock investing somewhat obscures that normal mechanism: Now, the highest goal of a typical investor is not so much a quarterly dividend check but rather a company whose value goes up 100x, so high-profile technology companies tend to spend their profits to become dominant players in artificial intelligence rather than just writing checks to shareholders. But most companies are not going to 100x their value; most companies make widgets and sell them for more than they cost to make and have some steady profits and think “well these profits are nice but they belong to our owners, so let’s give them to our owners.” Also, for various more or less good tax and corporate finance reasons, these days companies tend to give the money back to owners through stock buybacks (paying the shareholders for their shares) rather than dividends (writing checks to all shareholders). My two-step model makes sense to me, but you could have another model. You could have a one-step model something like this: - Companies try to maximize the cash that they send back to shareholders (through stock buybacks).
I don’t think that this is generally true about US public companies, but it is not a crazy model. Surely there are private companies like this; surely there are people who own small businesses (or large businesses for that matter) whose main desire is to get as much cash as possible every month, and who would much rather get big dividend checks than reinvest in growing their businesses. [2] Some people think that private equity firms tend to have similar priorities for their portfolio companies, that they load those companies up with debt and then prioritize current cash flow over reinvesting in the businesses. And I suppose you could have similar views about at least some public companies, something like “the executives get paid in stock, stock buybacks prop up the stock price, so the executives care more about short-term cash flows to shareholders than they do about long-term value creation.” I don’t really buy this — I think profits do more to prop up the price than stock buybacks do — but some people do. Anyway. If you are a populist politician, one thing that you might think is “Company X should lower its prices.” [3] How can you express that thought? What should you say to Company X to get it to lower its prices? Well, the simplest approach is just to say it: “Company X’s prices are too high and it should lower them.” Maybe you pass a law instituting price controls on Company X, saying it can’t charge more than $10 per widget or whatever. There are political and economic problems with this. Price controls get a bad rap, economically. Also Company X’s lobbyists are going to come to you and say “actually it costs us $10.15 to make a widget so your price controls will cause a collapse of the widget market and hurt the widget consumers you are trying to help.” Last week, President Donald Trump told credit card companies that they’re not allowed to charge more than 10% for their widgets, and the credit card companies gave the standard (and surely accurate) reply. But you might think that actually Company X is doing great and can easily afford to sell widgets cheaper. Perhaps you don’t want to meddle with price controls per se, but you want to at least shame Company X into lowering prices. One thing that you might say is: “Look, Company X’s profits are huge, surely it can afford to charge less for widgets.” If you are pretty draconian about it, you might say “we will pass a law saying that Company X’s profits can’t be more than _____.” [4] This law might have the effect of forcing Company X to lower its prices: If it is capped at a 10% profit margin, and it costs $9 to make a widget, Company X will have to charge at most $10 per widget to comply with the law. (The obvious workaround, though, is to increase costs, for instance by paying Company X’s executives a bonus of $1 per widget.) This is more or less how utility regulation works; it’s not so much the prices that are regulated but rather the return on equity. But, again, in the general case, there are political and economic problems with capping a company’s profits. In my little model, I assume that Company X is maximizing profits. But you might not assume that. You might have the one-step model, in which Company X is maximizing stock buybacks. If that is your model, you might try to force Company X to lower its prices by capping its buybacks. Same basic analysis: If Company X is not allowed to do stock buybacks, and it costs $9 to make a widget, and Company X has no use for its profits other than doing buybacks, then, without the ability to do buybacks, Company X will just say “meh I guess we’ll charge $9 per widget; we don’t have any use for any more money than that.” (Or various softer versions: Maybe Company X likes to keep $1 per widget in its cash reserve, but any more than that would go to buybacks or nothing.) I am making fun of this model here, but not entirely. It’s not completely nuts. If you were the owner of a company, and your company was allowed to make profits but not allowed to pay them out to you, you might care less about the profits. You might say “meh I’m not gonna get that money in my bank account this year so whatever, cut prices.” This is not a fully rational thought process — you might instead think “my net worth includes the value of the company I own, and the more profitable that company is the better off I am, even if I’m not getting a quarterly check” — but it is sort of psychologically plausible. [5] And the advantage of this approach is that, while economists are often (not always!) pretty skeptical of banning stock buybacks, it does seem politically more palatable. Politicians have been railing against stock buybacks for years; “people are worried about stock buybacks” is a long-running section heading in this column. Of course, there is also a history of politicians criticizing profits, but, you know, less. “Companies should not make profits” sounds a lot like socialism, and it does not go over well in US politics, certainly not in US Republican politics. “Companies should not do buybacks” is pretty mainstream and bipartisan. So “we will lower widget prices by forbidding widget companies from doing stock buybacks” strikes me as mostly incorrect economic analysis but plausible politics. Anyway here’s this: In an interview with The Wall Street Journal, President Trump’s Federal Housing Finance Agency director, Bill Pulte, questioned repurchases made by home builders when discussing the administration’s plans to lower housing costs. Pulte, the scion of a home-building giant, said the industry has purposely kept prices high, calling on it to help lower them and to build more homes across the country. “They’re making, in some cases, more money than they’ve ever made, and they’re buying back stock like never before,” Pulte said. “It’s something that we’re studying, you know, how much money they’re spending on buybacks.” He hinted at penalties for companies that didn’t help the administration’s push, adding that support for the government-backed mortgage-finance firms Fannie Mae and Freddie Mac shouldn’t be “used to fund buybacks at the expense of Americans who need to get in homes.” Sure. One novel and noteworthy feature of our current economic and cultural climate is that people love a scammer. Some of this is aesthetic: We love a rogue, someone who plays by her own rules and sticks it to the elites. But also, with crypto and prediction markets and meme stocks, we have developed better ways to profit from other people’s scams. If you know someone is running a pump and dump, the right move might not be “stay away” but rather “get in during the pump and get out before the dump.” It used to be that most people did not want to be involved in scams. Now, it seems, a lot of people are happy as long as they think they’re in on the scam. [6] This reverses a lot of traditional incentives. For instance: Traditionally, if you took $100 million of people’s money, and you never gave it back, and when they asked for it back you were like “teehee I know I’m such a rascal aren’t I,” you would have a hard time raising new money. There would be newspaper articles about your failed venture and your disgruntled investors, and if you wanted to find new investors for a new venture, you might have to change your name or move to another country or at least explain yourself. Whereas now, if you take $100 million of people’s money and never give it back and get stories written about you, you can go out to new investors and show them the stories and be like “lol look what a rascal I am,” and they will be like “oh man you are a rascal, here’s $200 million.” I realize this reads like a joke, and of course I am partly joking, but I’m also partly not. Adam Neumann’s shtick is not that he failed as the founder and chief executive officer of WeWork; his shtick is that he incinerated investors’ money in hilariously lavish ways while becoming a billionaire himself — and investors love that. We talked last week about the revival of MoviePass. We talked last year about the revival of Enron. I wrote: If you launched Enron Corp. on the stock exchange today — even without a business — the stock would go up, because that is funny. “Synonymous with willful corporate fraud and corruption” is the sort of meme that, in 2024, is valuable. Stocks go up in bankruptcy now. “The largest bankruptcy due specifically to fraud in United States history” is great!
Again, that sounds like I was kidding, and I was kind of kidding, and Enron was mostly kidding, but there’s also some truth there. Scandal, these days, sells. I mean it sells stock, even. Certainly it sells crypto. Donald Trump did some memecoins, made money for himself, lost money for investors, and keeps announcing new crypto things. “That guy did a crypto that took in a lot of money and then pancaked” seems to be kind of a good thing, in the memecoin world. The essential value of memecoins comes from attention, any kind of attention. Bad attention is fine. I once wrote about the memecoin launched by Haliey Welch, the woman who became famous for saying “hawk tuah” in a YouTube interview once: Yes I mean obviously Haliey Welch should get millions of dollars from selling a token representing her own virality, and obviously that token’s value should go to zero in like a day. That is how modern finance works! … People are complaining that this is a “rug pull” or a “pump and dump,” but I cannot understand what different thing they thought would happen. The Hawk Tuah coin would build an enduring business with large and growing value for years to come? What? Why? How? It’s a memecoin representing the fastest-melting imaginable form of fame; of course it should go to $500 million and then to zero in a day. That’s what it’s for! It worked perfectly!
My model here was “she is famous now, she won’t be famous later, you can monetize fame with a memecoin, so she should do that, and she did.” Fine. But an alternative model could be something like: “She is famous now, she won’t be famous later, unless she annoys everyone by doing a controversial memecoin that loses a bunch of money for people and gets her more negative attention. Then she’ll continue to be famous, or at least infamous, and maybe she can monetize that fame by doing another memecoin.” Here I am like 80% joking. Still, something to consider. Anyway: Eric Adams’ fondness for crypto — which during his one-term tenure as mayor of New York included the launch of the ill-fated ‘NYCCoin’ — is once again attracting controversy. Less than two weeks after being succeeded by Zohran Mamdani, Adams re-emerged to promote a new city-inspired cryptoasset called NYC Token. A portion of the proceeds would go to fight antisemitism and educate children about the blockchain, Adams said during appearances in Times Square and on the Fox Business television channel. Adams encouraged the public to buy into the project. The value of the token - which has no official affiliation with New York City or any state or local government - surged on Monday to as high as 58 cents before crashing within hours to under 10 cents. Crypto market observers and analysts have cried foul, drawing parallels to what’s known in the industry as a “rug pull” — crypto slang for a project in which the creators appear to withdraw or sell a significant number of tokens without explanation. “Our market maker made adjustments in an attempt to keep trading running smoothly, and as part of this process, moved liquidity,” according to an email statement sent from NYC Token. “The team has not sold any tokens and is subject to lockups and transfer restrictions.” … “It seems to be pretty obviously just another celebrity memecoin, which doesn’t really have any long-term value and just exists as a way to rinse retail traders,” said Tom Schmidt, general partner at venture capital firm Dragonfly. “Granted, if you’re choosing to trade Eric Adams’ memecoin, you probably know what you’re getting into.” Byrne Hobart calls this a “fame exit scam, where someone becomes a sufficiently high-profile public figure that they can launch a crypto token and then promptly sell a bunch of it”: The scam part isn't that the token dropped in value—these tokens generally don't promise any real connection between their price and some real-world source of value. Instead, the actual scam is reputational: someone gets famous enough to get away with it, and apparently decides that liquidating their fame at pennies on the dollar, something they'll be able to get away with exactly once, is worth doing.
And, yes, that — “something [he’ll] be able to get away with exactly once” — is my usual intuition. But, you know. He’s Eric Adams. When he comes to you and says “teehee I know I’m such a rascal aren’t I,” won’t you melt? Or at least laugh along with him? Donald Trump is proof that you can liquidate your fame at pennies on the dollar dozens of times, and each time it only gets more valuable. Perhaps doing a memecoin disaster as soon as he leaves office is the best way for Adams to launch the next phase of his career; perhaps this establishes his credentials. Luxury Retailer Saks Files Bankruptcy After Turnaround Fails. Blistering Metals Rally Sends Gold, Silver and Copper to Records. Netflix Weighs Amending Warner Bros. Bid to Make It All Cash. Wall Street Is Suddenly on the Defensive With the President. “The guy is a medieval prince. He needs to be buttered up.” McKinsey challenges graduates to use AI chatbot in recruitment overhaul. AI Startup Sued by Palantir Says Company Wants to ‘Scare Others’ From Leaving. “If I was flying business class, or flying somewhere where I’m representing somebody or something, I would think to dress more nicely.” Matthew McConaughey Trademarks Himself to Fight AI Misuse. 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! |
